Disclaimer:

This article is for general informational purposes as a high-level guidance for foreign investors exploring investments in Chinese startups. None of the information presented in this article constitutes any legal or regulatory advice. Readers should not act or refrain from acting on the basis of any information in this article without first seeking independent professional advice tailored to their specific situation. For clarifications on the exact investment and exit procedures, feel free to contact Haiwen & Partners or any other specialised law firms.

Summary:

• Foreigners can invest in Chinese startups that don’t fall into investors’ home country restrictions (export controls), as well as other applicable Chinese laws and regulations, such as the so-called Negative List or National Security Review criteria.

• Further stages are similar to such in Western countries: company-level approvals, government reviews for sensitive areas (similar to CFIUS in the US) and closing procedures.

• The main difference may be in applicable in capital controls and FX conversions that require additional procedures from the bank.

• Chinese startups can be divided in the following groups:

  • 1) the ones with overseas holdcos (e.g. Bytedance) that can be accessed directly at the offshore holdco level;

  • 2) domestic JVs that have both local and foreign shareholders and can accept USD investments (e.g. Unitree) — can also be invested directly – but at the onshore PRC level;

  • 3) domestic RMB-only companies that would not accept foreign investments, but could be accessed indirectly via RMB-denominated intermediary structures (QFLP) established in Mainland China and managed by a local GP.

• Offshore holdco is the most straightforward structure for both investment and exit, as the shareholding is built at the offshore level so it doesn’t involve proceeds repatriation from China at the time of the exit. The Chinese regulatory requirements also tend to be lighter. The trade-off, however, is a narrower range of investment targets.

• Onshore direct investment provides access to many Chinese deep tech startups, though domestic exits may require additional steps for repatriating proceeds to foreign investors.

• Onshore RMB structures (QFLP) offer the broadest coverage of investment targets and are well-suited as a long-term investment solution. While the initial setup requires more effort, the structure can be reused for future investments and provides a more streamlined path for exit repatriation.

“How can we invest in Chinese startups? Is it even legal?”

“How can we exit from our investment in China?”

“When we exit, how do we get our money back?”

These are the questions I often get even from professional investors who are just starting to explore VC investments in China.

Investing in Chinese startups is definitely legal. Moreover, Mainland China was the world’s 2nd largest destination of foreign direct investments (FDI) in 2024 (UNCTAD).

But there are a few aspects that a global investor should consider before investing in a Chinese startup:

• industry restrictions,

• investment structure,

• profit (i.e. exit proceeds) repatriation.

The first thing after identifying your potential investment target in China is to determine whether you as a foreigner are allowed to invest in this startup.

There are three main considerations that may limit your investment:

1. Restrictions from your home country. These are mainly export control rules. The most strict ones are imposed by the US government which restricts US citizens from investing in Chinese dual use, semiconductor, quantum and LLM companies (Outbound Investment Security Program). Export controls in the EU, UK and other Western countries apply mostly to investments in dual use companies that also includes tech transfer. It usually excludes passive minority investments.

2. China’s Negative List restrictions. Foreign investors are not allowed to own more than a certain percentage of companies in sensitive areas included in the so called Negative List issued by National Development and Reform Commission of the People’s Republic of China (NDRC) and Ministry of Commerce of the People’s Republic of China (MofCom). Some of the areas include:

   1. ICT: telecom, news, media, publishing — i.e. almost any digital platforms,

   2. Space: aerial mapping and imaging, air transport,

   3. Energy: nuclear power plants construction and operation,

   4. Biotech: human stem cells, genetic diagnosis.

   5. Full list is published here (in English).

3. China’s National Security Review criteria. If a foreign investor wishes to get either direct or indirect control or significant influence (e.g. board seat) over a Chinese company which is military-related or related to critical infrastructure (energy, food, IT, finance etc), you need to apply to NDRC for approval. Exact wording here (in English).

If none of the key restrictions apply, you may kick off the investment process.

Investment Process

Investment Process

After making an investment decision, there are four main stages of the investment process:

1. Company-level approvals (LOI/Term Sheet, SPA, SHA etc) + consents from the company’s directors and shareholders.

2. Filing procedures — most target companies only require a simple procedure of issuing register of shareholders, capital contribution certificates or share certificates and updating MofCom and SAMR registers and corporate information, which usually takes around a week.

   • China’s Governmental approvals — if the target company falls into the Negative List or in any other sensitive areas, the foreign investor and/or the target needs to obtain approvals from NDRC and/or MofCom (as applicable) and other authorities, such as Cybersecurity Administration of China (CAC) and State Administration for Market Regulation (SAMR).

3. Bank-level procedures related to China’s capital control rules and foreign currency conversion regulated by the State Administration of Foreign Exchange of the People’s Republic of China (SAFE) should also be considered.

Important to note that unwritten and informal requirements also apply broadly across each stage of the investment process. For example, in some cases you may satisfy all formal requirements, startup founders — especially the most popular ones — may still be reluctant from approving your investment because of various business, regulatory and/or timing considerations.

Despite these multiple layers of reviews, investment process in China is not that different from cross-border investments in Western deep tech startups:

• Industry-specific restrictions for foreign investors are imposed in most Western countries in areas like airlines, telecom, broadcasting etc.

• China’s National Security Review is very similar to the US CFIUS regulation both in terms of the filing procedure and reviewing criteria.

• Cybersecurity rules are also similar both in China and US/EU with GDPR and other data export restrictions.

• Same goes for anti-monopoly regulation.

Capital controls may be the main difference.

Investment Structures

From a corporate structure perspective, Chinese startups can be divided into two groups:

1. Companies with overseas holdcos (“Red Chips” or “P-Chips”) — “de jure” non-Chinese companies (Cayman, BVI, Hong Kong etc), but with significant operations in Mainland China. These operations are typically linked to the holdco in one of two ways:

   • WFOE: the foreign holdco directly or indirectly wholly owns the Chinese entity,

   • VIE: the offshore holdco indirectly holds a WFOE in the PRC and that WFOE obtains contractual control over the VIE company through a series of contractual arrangements. The VIE company itself is typically a purely domestic PRC company held by the founders or other PRC persons, which structure is commonly used in practice in sectors subject to foreign investment restrictions.

2. Domestic companies incorporated in Mainland China without any overseas holdcos. They are divided into two groups:

   • Joint Ventures (JVs) — companies with both PRC and foreign investors that are open to receiving foreign investments in USD directly.

   • RMB-only companies — companies backed exclusively by domestic investors. Although not restricted by the Negative List, they typically avoid foreign capital due to more complex closing procedures and increased regulatory scrutiny. They can accept only RMB investments.

Vast majority of Chinese startups belong to the “RMB-only” category, but most of them — except for the most sensitive ones — can be converted into JVs after admitting a foreign investor. Based on my analysis of cross-border deals, only 11% of investment rounds of Chinese startups in 2025 had at least one foreign investor, which can indirectly indicate an approximate share of JVs among all active startups in China. Only a fraction of all Chinese startups has overseas holdcos, although many of which may be more familiar to foreign investors such as Bytedance, Shein, DJI etc.

Based on each type of startup, various investment structures can be utilised:

1. If a startup has an overseas holdco, you can invest directly into that holdco entity in USD or other non-RMB currencies and not face any capital control issues at the holdco level.

2. If it’s a domestic JV company, you have to invest directly into its Mainland Chinese entity which involves the governmental review and capital control procedures. But you would become a direct shareholder of this startup.

3. If it’s a domestic RMB-only company and it’s not in the Negative List, the only way to invest might be through an onshore RMB structure (fund or single asset vehicle), although it is still subject to approval from the company. To achieve this, most foreign investors use the so-called QFLP structures, although legally you will still be treated as a foreign investor.

VIE structures were widely adopted by Chinese startups — including Alibaba, Tencent, ByteDance, and Shein — during the internet boom of the early 2000s. “Telecom value-added services” and “internet information services” fall under the Negative List as sectors where foreign investment is restricted or prohibited. VIE structures allowed internet startups to bypass those restrictions, raise capital from global investors, and list overseas.

Since 2020, China has introduced new rules that bring VIE structures under greater regulatory oversight. The Foreign Investment Law (2020) and the Trial Measures for the Administration of Overseas Issuance and Listing of Securities by Domestic Enterprises (2023) extended the scope of regulation to include “indirect” foreign investments and introduced obligatory filing of VIE structures with the CSRC (China’s equivalent of the US SEC).

In practice, the Negative List prohibits or restricts foreigners from directly investing in Chinese entities operating in restricted sectors such as certain specified internet services.

Qualified Foreign Limited Partner (QFLP) structure is a regulatory program that lets approved foreign investors commit capital into RMB-denominated PE/VC funds (multi or single asset) established in China. In practice, a foreign LP (or group of foreign LPs) is licensed/qualified under a local QFLP pilot, converts foreign currency into RMB under agreed rules, and then invests through a domestic vehicle managed by a local licensed GP.

After obtaining all licenses and qualifications, the QFLP structure might simplify the investment and repatriation processes for all consequent deals made through the same vehicle, although the first transaction would require even more administrative efforts than a direct investment into an onshore company.

The exact QFLP configurations vary across different city/province-level pilot regimes (e.g., Shanghai, Beijing, Shenzhen, etc.). You can find here a more detailed description of the Shanghai QFLP program.

The investment process is also slightly different for each structure:

1. Offshore holdco is often the most straightforward structure, as it involves fewer reviews on the China side, unless the investment resulted in the change of the control.

2. Onshore holdco can involve more PRC regulatory procedures and therefore tends to require the most coordination.

3. QFLP involves the same level of PRC regulatory procedures as onshore holdcos, but it simplifies the bank procedures for all consequent deals and might potentially provide an access to a larger pool of RMB-only companies.

Exit process

Investing in a Chinese startup may not be as hard as exiting from it.

There are three main hurdles for exiting from a Chinese startup, some of which may be by reason of business opportunity and market consideration:

1. Exit opportunities — IPOs, M&As, secondary private placements.

2. Government reviews — Negative List, National Security and anti-monopoly regulation.

3. Capital control rules — profit (i.e. exit proceeds) repatriation.

Below we’ll dive into each of these hurdles.

Exit opportunities

In 2025, China was the second largest VC/PE liquidity market after the US and the trend is positive (more on this in my previous articles here and here).

There are three main exit scenarios from Chinese startups:

1. IPO is the most common exit path in China which is available for all types of startups: domestic ones (Shanghai, Shenzhen, Beijing) for local companies, offshore ones, like Hong Kong (HKEX) or even US (NASDAQ), for the FIEs and overseas holdcos.

2. M&A is a less common exit path for Chinese startups. Domestic M&A are more frequent, but usually smaller in sizes. Cross-border tech M&As are rare, so sticking to onshore structures makes it easier to find a potential acquirer.

3. Minority stake sale in the secondary market, depending on the specific company, industry and/or foreign investor(s) involved, tends to be the easiest exit route for the same company from a commercial perspective, compared with an IPO or M&A, if it does not involve a change of control or influence that could trigger additional scrutiny. That said, each transaction should still be assessed on a case-by-case basis.

Government reviews

These restrictions apply equally to all investment structures: onshore companies and offshore companies with a significant presence in China.

However, they vary across different exit strategies.

Cross-border M&As are the most complex, as they may involve all kinds of security-related governmental reviews depending on both the nature of the acquisition target and the acquirer itself.

An example of this issue may be the Meta acquiring Manus case:

• Although Manus is a Singapore-based company, most of its operations have been originated from Mainland China.

• In early 2025, the US VC firm Benchmark led the company’s Series B, which hasn’t triggered any significant scrutiny from China, because Benchmark only acquired less than 15% of Manus.

• However, when Meta acquires Manus, the change of control prompted scrutiny of export control and/or National Security and Cybersecurity reviews.

IPOs are generally more complex than M&As due to standard disclosure procedures. The most pressing issue with overseas IPOs is usually related to sensitive data export, especially for large consumer platforms, e.g. Didi. Meanwhile, China’s leading LLM player Minimax with ~60 million users in China successfully went public on HKEX in 2025.

A minority stake sale in the secondary market, depending on the specific company, industry and/or foreign investor(s) involved, tends to be generally the easiest exit route for the same company, compared with an IPO or M&A, if it does not involve a change of control or influence that could trigger additional scrutiny. That said, each transaction should still be assessed on a case-by-case basis.

Generally, any cross-border and overseas exits tend to trigger much heavier government scrutiny than domestic exits.

Capital controls & taxation

Capital controls are much easier for cross-border and overseas exits than for pure domestic cases. But different investment structures involve various level of repatriation complexity.

Offshore holdco is the easiest exit structure from the perspective of repatriating the exit proceeds, because your shares are held in an offshore jurisdiction usually with no capital control rules.

• If the company is acquired, the acquirer usually can transfer your money directly from a dedicated overseas bank account.

• If the startup goes public outside of Mainland China, you can sell your shares on the global public markets after the lock-up period.

• The only complication may arise when the global entity goes public in Mainland China, which is rare. In this case, after the expiration of the lock-up period you can sell your public equities as a pre-IPO investor on the onshore stock exchange without any additional license (e.g. QFII) if the transaction is only for the exit purpose. After that the proceeds first go to your dedicated RMB account. Then after settling the withholding on the corporate income tax (10% for non-residents) and undergoing a bank review, the exit proceeds can be transferred to your overseas account.

• Worth noting that under China’s tax regulation, exiting from overseas companies may be considered as an overseas indirect transfer if the company’s actual business originates mostly from China. Under Bulletin 7 and relevant rules, such an indirect transfer of PRC assets by a non-PRC tax resident may be re-characterized as a direct transfer of PRC taxable assets if the transaction lacks reasonable commercial purposes. In this case, PRC-sourced gains from the transfer may be subject to enterprise income tax, and tax filing and withholding obligations may be triggered — effectively making the deal taxable in the same way as exiting from a domestic Chinese company.

Domestic holdcos (both direct and via QFLP) require more administrative steps for repatriation, as investors need to complete bank documentation reviews for each transaction before transferring the exit proceeds out of China — a process that, while standard, can add time and complexity compared to offshore structures.

• If your shares are acquired by a foreign buyer with an overseas bank account, it is theoretically possible to request the buyer to transfer the exit proceeds directly to your overseas bank account to avoid repatriation issues. However, this may still require paying the withholding on the corporate income tax which must be settled onshore in RMB.

• In case of IPO, same as for overseas holdocs, if the company is going public overseas, you can sell your shares after the lock-up period and transfer money directly to your account.

• If the company goes public in Mainland China or if it’s acquired by a local buyer, you need to file repatriation of your proceeds with the Chinese bank. For this, the bank will review transaction documents, including source of income, tax settlement confirmation, your foreign company information etc. This process is also common for any exit from investment in Western countries, with the exception that no tax clearance confirmation is required.

Onshore QFLP structures involve more upfront steps than investing directly into a Chinese holdco, as both the foreign investors (LPs) and the local GP need to meet additional qualification requirements. However, once the QFLP is established, it may simplifies the repatriation process of exit proceeds for all consequent deals made through the same licensed vehicle.

Below is the summary of the step-by-step process of repatriating exit proceeds for foreign investors from selling shares in Chinese onshore entities, based on the guide of K-Bright Law Firm (in Chinese).

Under Article 22 of the Implementing Regulations to the Foreign Investment Law (外商投资法实施条例):

“A foreign investor’s capital contributions made within China, profits, capital gains, proceeds from asset disposals, royalties from licensed intellectual property, compensation or damages lawfully obtained, liquidation proceeds, and the like may be freely remitted into or out of China in RMB or foreign currency in accordance with the law. No entity or individual may illegally restrict the currency, amount, or frequency of such inward or outward remittances.”

In principle, there is no legal barrier preventing a foreign investor from exiting through an equity transfer and remitting its investment returns offshore. In practice, however, the operational requirements of different regulators, complex documentation and various bank review standards require detailed arrangements to reduce potential risks.

This is why engaging a specialised legal advisor during the repatriation of exit proceeds is crucial.

Final remarks: which structure is better?

Everything depends on the goals of the investors:

• If you want to make investments in China, offshore holdcos might be the best choice, because they are the easiest to start with, no capital control issues, most target companies are global in nature and well-knows outside of China. They also may be more open to foreign investors. However, the choice of available opportunities is quite limited.

• If you are interested in a specific startup that doesn’t have an overseas holdco, investing directly into an onshore holdco can be an option, but may be subject to longer repatriation process with stricter requirements from the banks. In this case, securing a good Chinese tax or legal advisor is always a good idea.

• If you are thinking about a broader long-term strategy of investing in Chinese startups, creating your own QFLP structure may be a preferred option, as it might give you higher coverage of investments with potentially more simplified investment and exit procedures for all consequent deals made through the same vehicle.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Summary:

• Ten robotics segments are already VC-scalable: autonomous driving, humanoids & animal robots, drones & marine robots, medical robots, service robots, warehouse robots, hardware components, software for robots, robotic arms, laboratory robots.

• VC-scalability varies across countries, e.g. hardware and service robots are easier to scale in China, while pure software plays are more sustainable in US/EU.

• China is leading by the number of unicorns and VC/PE exits in robotics, while US has the largest VC-backed robotic companies (Intuitive, Waymo, Figure, Anduril).

• The following five areas have the largest VC-upside in the next 5-10 years: autonomous driving, drones & marine robots, warehouse robots, medical robots and service robots.

• Autonomous driving will probably be the first segment to produce a $1T robotic company in the next 10 years, while it will take longer for humanoids and medical robots.

• AI & Software is a critical part for the robotics progress, but the business model of pure software players is hardly sustainable in robotics in the long run.

• China will keep dominating the robotic supply chain, both hardware and software, although software will most likely be open source and hard to monetise.

• Western players will be much stronger in applications and vertical solutions due to their proximity to large and mature markets.

For the past decades, the world has been shaped by digital innovation: first the internet, then mobile and cloud, and most recently — AI.

These technologies gave birth to the world’s first $1T+ companies (Facebook, Amazon etc) and emerging $100B+ unicorns (OpenAI, Databricks, Anthropic, Bytedance), nurtured by the top tier VC firms like Sequoia, Kleiner Perkins, a16z and others.

However, the contribution of software to the world’s GDP is still low, e.g. only 4% in the US. The majority of global value creation still remains in physical industries, such as manufacturing, construction, logistics. All these sectors are built on physical interaction and have historically been insulated from pure software disruption.

Robotics was meant to transform these physical industries, but almost none of the robotics pioneers managed to reach even $100B in valuation:

• Unimation, the world’s first robotics company — acquired by Westinghouse for $107M in 1983.

• KUKA developed their first robot in 1973 — acquired by Chinese company Midea for €4.5B in 2016.

• ABB Robotics, world’s largest industrial robotics manufacturer — acquired by Softbank Robotics for $5.4B in 2025.

• EPSON, the world’s 2nd largest industrial robotics manufacturer — less than $4B market cap as of Nov 2025.

• Boston Dynamics, the pioneer in humanoid and animal robotics — acquired by Google, then by Softbank, then by Hyundai. The last price was $1.1B in 2021.

Moreover, none of these companies were backed by VCs.

Why has robotics been always hard for VCs?

The short answer — scalability. Hardware businesses are less scalable than software because you need to procure components, produce and assemble them and then move physical products across sophisticated supply chains to the other part of the world.

Niche applications is another even bigger issue. In areas like automotive manufacturing, semiconductor fabrication and packaging, warehouse fulfilment automation already reaches 70-90%. However, deploying robotic solutions at every facility requires high customisation and support from system integrators, which significantly limits scalability.

All these issues make it hard for VCs to underwrite a potential 100x return from robotics investments.

Physical AI — A New Hope

This may be starting to change now.

Over the last few years, hardware innovation is getting back on the radars with NVIDIA’s Jensen Huang predicting that “the next wave of AI will be Physical AI”, referring to AI penetrating the physical dimension in a form of more advanced robots, autonomous driving, drones and other smart devices.

The timing seems to be right.

Demand side — aging population and labor shortages in developed countries:

• By 2030, one in six people globally will be over 60.

• By 2050 the population aged 80+ will nearly triple to around 425M.

• 0.5–1.5% annual workforce decline in Japan, South Korea, Italy, Germany, China.

• US manufacturing industry is projected to have 2.1M unfilled jobs by 2030.

• Logistics and warehousing report 40–60% vacancy rates during peak periods.

• Construction in Europe and Japan has 20–30% of roles consistently unfilled.

Supply side — progress in AI and reducing hardware costs.

• Large-scale multimodal models and RL-based control systems improve perception, dexterity, and generalisation.

• Component prices (servo motors, sensors, batteries, compute) have dropped 30–70% over the past decade due to Asian supply chains and Chinese scale manufacturing.

All this is resulting in bullish forecasts by the leading consulting firms predicting the physical AI and robotics market to reach $300B by 2030 and $1T by 2040-2050.

VC-Scalable Segments in Robotics

To identify which robotics areas have the highest potential to reach a VC scale, I’ve collected data on 123 publicly known VC-backed robotics companies with $1B+ valuation, including unicorns, public companies and acquired companies. The exact methodology is described at the end of the article.

This analysis does not only pinpoints the most hyped areas, but also highlights the following key considerations about these sectors:

• Market size. Investors understand that the potential market is big enough to reach at least a $1B+ outcome (VC-scale).

• Scalability. Investors expect that companies in this segment can grow exponentially every year without high customisation for each deployment site.

• Moat. Investors believe that due to the economies of scale and the ecosystem flywheel these companies will be defensible and the barriers will grow, preventing new competitors from entering the space and compressing margins.

• Access to capital. As a consequence of the previous three points, investors are ready to deploy more capital into these companies, which will help them scale faster and become more defensible in the future.

The abundance of $1B+ VC-backed companies in certain segments also proves that early backers of those startups can return at least 10-20x from their investments.

Based on this analysis, I’ve identified the ten robotics segments that already indicate VC-scalable growth.

1. VC-hyped segments:

• autonomous driving (including robotaxis, autonomous trucks),

• general-purpose bio-inspired robots (humanoids, robot dogs etc),

• unmanned aerial & marine systems (drones, vessels, eVTOLs).

These segments have the biggest number of unicorns, while not as many exits yet. Valuations are high, but the actual market adoption is nascent, as positive ROI from implementing those technologies is still unreachable. Nevertheless, VC investors are actively pursuing those opportunities, as they promise the next generation of general-purpose hardware platforms, akin to cars in the early 20th century.

2. Mature VC-scalable segments:

• medical robots (first of all surgical robots),

• non-humanoid service robots (cleaning robots, vacuum cleaners, operators with hybrid robotic platforms),

• warehouse automation (AGVs/AMRs and other robotics solutions),

• hardware components (mostly LIDARs and edge compute chips).

Most of these segments already have significant amount of $1B+ acquisitions or IPOs, current market sizes are measured in dozens of billions of dollars with real adoption and proven positive ROI.

3. New VC segments

• pure software for robotics (robotic brains and traditional SLAM),

• general-purpose robotic arms (adaptive or collaborative arms),

• laboratory automation (robotic solutions for autonomous labs).

These segments don’t have many sizeable exits yet, but there are some unicorns, which proves that VC-scale is possible.

4. Missing segments

• Automated production lines — there are incumbent players like ABB, FANUC, Siemens etc, but no sizeable VC-backed unicorns, because these solutions require a lot of customisation and integration support, which hinders scalability.

• Basic robotic arms — market is commoditised with hundreds of producers and no clear product differentiation.

• Basic hardware components (actuators, electric motors, gearboxes, drives) — same as for generic robotic arms: commoditisation and no defensibility.

• Pure integrators — consulting companies that deploy robots on the customer premises; although being a crucial link in the chain, these players are rarely VC-backable due to their service-oriented revenue and low scalability.

Which Countries Are Leading?

Out of 123 unicorns and $1B+ exited companies:

• China — 68 companies (55% out of total),

• USA — 39 companies (32% out of total),

• Europe — 16 companies (13% out of total), with Israel and Germany as leaders.

Nevertheless, out of the TOP 5 companies by valuation / market cap, four companies are American and only one is Chinese — DJI.

TOP 10 VC/PE-backed robotics companies:

1. 🇺🇸 Intuitive Surgical — $200B+ (public)

2. 🇺🇸 Waymo — $45B

3. 🇺🇸 Figure — $39B

4. 🇺🇸 Anduril — $30.5B

5. 🇨🇳 DJI — $20B

6. 🇨🇳 Yinwang / Huawei — $18B

7. 🇨🇭 Aptiv — $17B

8. 🇨🇳 Horizon Robotics — $14B (public)

9. 🇨🇳 Zongmu — $11.4B

10. 🇮🇱 Mobileye — $9.2B (public)

$1B+ robotics exits:

• China — 17, mostly IPOs

• USA — 16, both M&As and IPOs

• Europe — 7, mostly IPOs

Historically, the USA have been the largest exit market for robotics companies, but last year China caught up with four new robotics IPOs in 2025 (Geek+, Seyond, CiDi, Breton) and five IPOs in 2024 (Horizon Robotics, Robosense, Weride, Pony.ai, Dobot).

Competitive dynamics varies across each segment.

Americans are leading in:

• Medical robotics — with an ultra success story of Intuitive Surgical.

• Robotics software — Skild AI, Physical Intelligence, Field AI that are VC-backed + other leading labs like Google DeepMind (RT-2, RT-X), OpenAI (Aiko), NVIDIA (GR00T)

• Laboratory automation — with Opentrons, Equashield, and Periodic Labs.

Chinese are leading in:

• Service robotics — among $1B+ VC-backed cleaning-related players, all companies are Chinese, especially after the US consumer robotics pioneer iRobot went bankrupt in Dec 2025.

• Autonomous vehicles — despite the growing adoption of Waymo’s robotaxis, there are dozens of Chinese unicorns targeting passenger, commercial and industrial segments.

• Hardware components — although NVIDIA Jetson chips are a household name in global robotics, it only contributes to ~1% of NVIDIA’s revenue; meanwhile, China is the absolutely leader in LIDARs (Hesai, Seyond etc) and is catching up with its edge compute chips.

• Warehouse automation — although the legacy players like Vanderlande and Dematic are American and European, many major next generation VC/PE-backed companies are Chinese, e.g. Geek+, Siasun, HAI Robotics. American company Symbotic is another significant player ($36B market cap), but it’s not VC-backed.

Equal competition:

• Unmanned aerial & marine systems — although DJI is the market leader in standardised commercial drones, there are several specialised American and European startups, especially in defence that might not have the same revenue scale as DJI, but are valued at much higher multiples, as they provide integrated solutions for the end customers.

• Humanoids & animal robots — although China is dominating hardware supply chains for humanoids and robotic dogs (Unitree, UBTECH etc), the US and European players — like Figure, Optimus, 1X — are still much stronger in AI capabilities, although China is catching up.

One of the conclusions is that VC scalability varies not only across different segments, but also across different ecosystems. Manufacturers of hardware components and service robots — especially cleaning robots — are much harder to scale in the West, than in China. Likewise, pure robotic software developers are more sustainable in the US than in China where software is harder to monetise.

When Will We See a $1T+ Robotic Company?

Robotics automation has existed for decades, but the mature segments haven’t yet shown the scalability needed to produce a $1T company — unlike AI. As of now, the largest robotics company is the medical robotics pioneer Intuitive Surgical with a ~$200B market cap.

Below I’ll try to summarise my estimations on the upside potential for VC-backed companies in each robotics segment across different time horizons:

• 5 years (by 2030) — for VC/PE growth funds,

• 10 years (by 2035) — for early-stage VC funds,

• 25 years (2050 and beyond) — for strategic investors.

These estimations are based on assumptions across the following factors:

• Three VC-scalable factors: total addressable market (TAM), scalability and moat;

• Size of the existing incumbents that the new companies are disrupting;

• Size of major companies in other industries with similar business models;

• Speed of disruption and stickiness of the existing incumbents.

The upside potential doesn’t always progress linearly over time; it depends on where the technology sits on the Gartner Hype Cycle. For example, the rise of physical AI may drive near-term valuation uplift for humanoids, but since large-scale adoption is still at least a decade away, their mid-term upside doesn’t increase meaningfully.

Autonomous Vehicles (AVs)

The sector is shifting from localised pilot programs to a global infrastructure utility with a Mobility-as-a-Service (MaaS) model where autonomy is embedded into public infrastructure, logistics and industrial fleets.

By 2030: Further Robotaxi Roll-outs ($100B+ potential)

TAM is still relatively small, but expanding fast as robotaxis scale in dense cities and autonomous trucking is adopted in fixed corridors and private industrial domains. Waymo is already valued at ~$45B in its 2024 round and is reportedly discussing a new raise at $100B+, while Chinese leaders (e.g., Pony.ai, WeRide) are also ramping fleet scale globally. Scalability remains moderate because fragmented regulations, safety cases, edge conditions, and “city-by-city” localisation keep rollouts constrained.

2030–2035: Robotaxis & Commercial Fleet Adoption ($500B+ potential)

Robotaxi operators should capture the bulk of value as they replace traditional ride-hailing services like Uber, and could surpass incumbents to reach $500B+ market caps as removing human drivers increases take-rate and improves unit economics. Autonomous long-haul trucking, industrial vehicles, and public transport should also see meaningful adoption by 2035, but those markets are either too niche or too fragmented — with weaker moats — to produce single players at $100B+ scale. Logistics companies (e.g., UPS $85B, FedEx $70B, DHL $60B) may capture major value by deploying autonomous rides to cut costs and increase capacity, yet their market caps will likely still be in the ~$300B range. In industrial trucks (e.g., mining/construction), most value is captured by incumbent OEMs like Caterpillar ($270B+ market cap), so the largest upside requires directly disrupting those players.

After 2035: Mass Consumer Adoption ($1T+ potential)

Over time, new entrants may still capture significant share in autonomous commercial vehicles, but the largest upside should come from automating consumer passenger mobility. This likely happens either through new and legacy automotive OEMs adopting ADAS at scale, or through robotaxis becoming a cheap substitute for individual car ownership. In an ADAS-led path, competition will eventually compress margins and cap valuations near today’s automotive OEM levels (~$500B) or around $1T similar to Tesla. If robotaxis become a true alternative to owning a car, operators can capture far larger value ($1T+) by controlling the full passenger-mobility loop, with strong moats from decades of safety data and tightly integrated hardware/software supply chains. Other non-obvious winners could be consolidating providers of data infrastructure — safety assurance/validation, simulation + synthetic data pipelines, mapping/localisation, fleet orchestration.

Humanoids & Animal Robots

This segment is transforming the whole robotics industry from specialised robots into general-purpose hardware platforms replacing / augmenting manual labor in all areas.

By 2030: R&D and Niche Deployment ($10B+ potential)

The market is still small as the industry remains in a heavy R&D and product development stage, with adoption limited to pilot programs in automotive plants and logistics centers. Although Figure has reached $39B valuation in its Series C and it may well reach a $100B+ valuation, it’s hardly sustainable. Scalability is constrained by high BOM + reliability + service costs, while robodogs scale earlier in inspection/defense but remain niche with capped TAM. Moats are either in large-scale / low cost hardware capabilities (e.g., Unitree with $4B valuation) or in proprietary datasets and AI models and quick real-world deployment feedback loops.

2030–2035: Car-Like Adoption ($100B+ upside)

After 2030, we may see the first real humanoid deployments that directly replace or augment manual labor in industrial and commercial facilities (manufacturing, warehouses, retail, etc.) that are not yet fully automated. This adoption wave may mirror the rise of cars in the 20th century: a split between robotic OEMs (general-purpose hardware platforms with basic intelligence) and robotic integrators that build their own AI models and sit closer to distribution, but remain tied to specific use cases. Hardware OEMs will be more scalable yet eventually commoditised, so their upside may resemble today’s automotive manufacturers (~$200B). Robotic integrators may also top out around $100–200B because, despite a large market opportunity, the model remains service-heavy and harder to scale; animal robots follow a similar pattern at smaller scale in niches like inspection. In this period, none of these systems are truly general-purpose — each still needs adaptation to specific tasks.

After 2035: Smartphone-Like Adoption ($1T+ upside)

As with AVs, the largest upside for humanoids is consumer adoption. To be useful in unpredictable household environments, robots will need highly sophisticated AI plus adaptivity and continuous learning—initially enabled by teleoperation and ultimately reaching full autonomy. Once that threshold is crossed, adoption could become genuinely general-purpose, akin to smartphones, with humanoids serving as universal physical companions for a household or even each individual. That implies a massive $5T+ TAM and creates room for Apple-like winners that integrate hardware + software supply chains, distribution, and ecosystem lock-in. Such platforms could plausibly reach $1T+ market caps.

Unmanned Aerial & Marine Platforms

The unmanned aerial & marine autonomy stack is evolving from “tools” into software-defined infrastructure for defence, logistics, offshore energy, and (eventually) passenger mobility.

By 2030: Defence Driver ($100B+ potential)

Commercial adoption is proven, but it’s driven mainly by defence, inspection and mapping. Scalability is high for standardised hardware producers, such as DJI ($20B valuation) that produce drones at mass-scale (70% of the global market) — their moat is in the economies of scale. Vertical software-led players have even stronger moats, because they own industry-specific software and autonomy stack, along with the customer relationships. But because of the tight airspace regulation and early stage of the technology, in the near term only the defence-focused players like Anduril ($30.5B valuation) are likely to gain significant scale — similar to incumbents like Lockheed Martin ($135B), Northrop Grumman ($95B).

2030–2035: Vertical Use Cases (~$200B potential)

Use cases shift from defence to civil industries, led by drone logistics (campuses, industrial sites, offshore assets, and corridor-based routes), energy and utilities inspection at scale (grids, wind/solar, pipelines, offshore), and autonomous maritime operations (ports, coastal shipping, offshore support, subsea inspection). However, their scale will be capped by $100B, similar to existing incumbents like FedEx ($70B), Maersk ($35B), Quanta Services ($70B). Horizontal drone OEMs will continue to scale, but limited product defensibility is likely to compress margins as vertical players increasingly internalise hardware. Urban aerial mobility will see the first meaningful pilots (eVTOLs, drone delivery), though deployment is limited by airspace regulation and local noise & safety permitting.

After 2035: Mass Vertical Adoption + Early Urban Mobility ($200B+ potential)

The markets for drones and unmanned maritime vessels in logistics, inspection, and offshore operations are likely to mature and consolidate with market caps hardly exceeding $200B. EVTOL deployment will be accelerating, providing faster alternatives to car ride-hailing or even short-term airline flights. This adoption will most likely be concentrated in public and shared transport systems, rather than mass-consumer ownership, which will cap the scale of the eVTOL operators roughly at the size of the existing ride-hailing companies like UBER ($200B) or Waymo ($45B).

Medical Robotics

Surgical robots are by far the largest segment of the medical robotics market which is evolving from complex robotic tools towards modular, affordable and AI-powered platforms enabling co-piloted and tele-operated procedures. All other segments (rehabilitation robots, exoskeletons etc) will stay niche, while non-surgical robotics assistants are included in the humanoid and service robot sections.

By 2030: David vs Goliath (<$20B potential)

Surgical robotics is dominated by a single player — Intuitive Surgical ($200B+ market cap) that is growing ~20% YoY with ~30% net margins. While the core surgical robots market appears niche, it is sizeable ($30–50B TAM) and still early, with only ~10–15% penetration. Importantly, robots account for just ~25% of Intuitive’s revenue, with the majority coming from high-margin, recurring instruments and service — materially expanding the effective TAM. Scalability is currently gated by long hospital CAPEX cycles and intense regulatory requirements, while the moat is in high switching costs and specialised surgeon training. New modular entrants like CMR Surgical are expanding the market by targeting “new-to-robotics” sites with lower pricing, but their near-term valuations are likely to remain in the $20B range.

2030–2035: Gradually Disrupting Intuitive ($100B+ potential)

We may expect a further disruption of Intuitive, with the emerging players winning over an increasingly larger market share due to their modular and more affordable system designs, as well as AI co-pilots and remote collaboration features. Although Intuitive also heavily invests in these technologies, the gap between legacy and new players will narrow. Intuitive itself might reach a $1T market cap by now, while its growth will mostly likely be stagnating.

After 2035: Oligopoly ($300B+ potential)

Next-gen surgical robotics platforms will keep pressuring Intuitive’s dominance and could plausibly scale to $300B+ market caps — unless Intuitive or other incumbents acquire them first. Even in that scenario, Intuitive is likely to remain a top-tier winner given its installed base, consumables flywheel, and regulatory “safety trust” moat. Exoskeletons and most non-surgical medical robots should grow, but will likely stay niche relative to the surgical platform opportunity.

Non-Humanoid Service Robots

These types of task-specific robots will eventually disrupt multiple traditional industries and give birth to a next generation of operating companies, although at some point it may face some competition with general-purpose humanoid robots.

By 2030: First Mature Niches ($30B+ potential)

The sanitation and cleaning industry is the early adopter: cleaning robots and autonomous lawnmowers are already becoming a mature niche with established players (Ecovacs, Dreame, Keenon, etc.). Standardized customer requirements across industries, a sizeable market, and low regulatory hurdles enabled smooth scaling — although commoditisation will cap upside (typically ~$20–30B market caps), making funding and brand awareness key success factors. Meanwhile, robotics is emerging as a core differentiator in other niches — agriculture (XAG), predictive maintenance (Gecko Robotics), construction (Icon) — but these remain early because the tech is still nascent and market education is required.

2030–2035: Rise of Robotics-Powered Operators ($100B+ potential)

As robots get more capable and affordable, traditional industries increasingly deploy robotic assistants to complement human labor — e.g., autonomous carts for last-mile and indoor delivery in hospitals and hotels, harvesting robots in agriculture, and kitchen/waitress robots in F&B. This will create two types of companies: service-robot suppliers and robotics-powered operators. Suppliers may scale fast, but commoditisation and intense competition likely keep their valuation upside in the ~$50B range, similar to the current electronics manufacturers; operators, while less scalable, can capture more value by disrupting incumbents (McDonalds, Doordash, Cargill, etc.) and may eventually reach $100B+ market caps. via structurally better unit economics, although it may require decades.

After 2035: Task-Specific vs General-Purpose (~$300B potential)

By 2035, non-humanoid service robots will likely be embedded across many industries, supported by a next generation of operating companies that have matured into large-scale businesses. After 2035, humanoid robots may finally reach the technological threshold to meaningfully replace or augment human labor across a wide range of use cases. Even then, general-purpose humanoids will likely co-exist with task-specific non-humanoid robots: humanoids handle flexible, unpredictable tasks typically done by humans, while specialized robots remain the most efficient for repeatable workflows. Human labor shifts toward managing and maintaining robots, plus higher-level tasks and decision-making. VC-scalable robotic operators will be gradually replacing traditional operators in each niches and better unit economics will provide potentially higher upsides (~$300B).

Warehouse Automation

The sector is evolving from rigid, hardware-heavy conveyor systems and “islands of automation” into fully autonomous “dark warehouses” with an AI-orchestrated ecosystem of mobile robots and intelligent manipulators.

By 2030: Disrupting Legacy Players ($50B+ potential)

The market is already sizeable as legacy integrators like Vanderlande, Dematic, and Daifuku maintain $3-4B revenues, while the public non-VC-backed player Symbotic shows the market’s massive upside with a $36B market cap and 20–30% growth. Scalability is driven by the global rollout of proven AS/RS and AMR stacks, though challenges remain in integrating these with legacy warehouse management systems (WMS). Moats are shifting from pure mechanical engineering to software-driven orchestration and workflow lock-in, with VC-backed challengers like Geek+, Exotec, and Fabric poised to capture significant market share and reach $30-50B valuations.

2030–2035: Rise of “Autonomous Warehouses” ($100B+ potential)

The market growth is driven by the adoption of “autonomous warehouse” ecosystems with true piece-picking at scale via tactile-aware manipulation and robotics foundation models for open-ended “physical work.” The standard adoption of autonomous truck loading/unloading modules and warehouse “autopilot” layers that reduce the need for manual intervention. Scalability improves through modular automation that speeds up multi-site replication across different building types. Moat concentrates in whoever owns end-to-end orchestration + performance data, enabling a shift from equipment projects to Labor-as-a-Service economics and creating room for new “warehouse OS” entrants alongside incumbents.

After 2035: Mass Adoption ($300B+ potential)

The “lights-out” warehouse becomes the global standard, with growth driven by penetration into smaller facilities and emerging markets that were previously uneconomic to automate. Scalability reaches its peak through standardised hardware platforms and autonomous exception handling, allowing for 24/7 operation with minimal human oversight. The long-term moat hardens into an oligopoly built on ecosystem lock-in. By 2050, the leading automation platforms will function as the mandatory “operating systems” for the global supply chain.

Hardware Components

This segment is foundational to robotics, but early commoditisation and weak defensibility will cap upside — unless a few players create durable moats via platform lock-in, standards, or ecosystem software.

By 2030: Premature Commoditisation ($10B+ potential)

By 2025, much of the robotics BoM (actuators, motors, drives, batteries, sensors) already has dozens of largely interchangeable Asian suppliers, so value capture is thin and few VC-scale winners emerge. LiDAR was the exception because it was technically hard, but it is rapidly commoditising as Chinese leaders (e.g., Hesai, RoboSense) out-scale Western peers, due to lower costs and massive economies of scale. These companies have already reached multi-billion dollar valuations and can still compound 2–3x, but persistent pressure from new competitors and price curves keep the moat fragile.

2030–2035: Scale-Up Phase (~$100B potential)

As robot volumes grow, component suppliers scale revenues, and the market consolidates around the best-funded players with manufacturing scale, quality/reliability data, and distribution. However, differentiation remains limited for most SKUs, so winners look more like best-in-class industrial component companies than software monopolies. The top few platforms could reach ~$100B market caps — comparable to the current manufacturers of hardware components for industrial automation, such as Parker-Hannifin ($112B), Keyence ($83B), Emerson ($77B).

After 2035: Search for a Moat ($300B+ potential)

Commoditisation remains the default; the only path to outsized $1T+ outcomes is owning a true bottleneck and wrapping it in a sticky software layer (toolchains, safety certification, reference designs, developer ecosystem) — a “NVIDIA’s CUDA-like” lock-in for robotics. The challenge is that robotics supply chains are fragmented, with no single chokepoint like datacenter GPUs, so ecosystem plays are hard and even more capital-intensive. Edge compute platforms or developer boards are the most plausible wedge and some companies are already pushing this narrative (like D-Robotics, less than $1B valuation yet), but it requires massive funding, standards adoption, and a defensible developer flywheel rather than just better hardware.

Software for Robots

The first wave of robotics software startups in the 2010s focused on SLAM and navigation, but they were mostly project-based, tightly coupled to specific robot platforms, and hard to scale across environments. As a result, most pioneers (e.g., Covariant, 6D.ai, Metaio) were acquired before reaching meaningful scale.

By 2030: The Next Hype ($50B+ potential)

In the 2020s, foundation models spawned a new cohort — Skild AI, Physical Intelligence, and Field AI — already at $1B+ valuations. Because the “robotic brain” is the key bottleneck for useful general-purpose robots, investors frame these as the next OpenAI/Anthropic for robotics. But they were founded only ~2 years ago and the trend is very recent, so near-term valuations will be driven more by demos than business metrics; upside could reach ~$50-100B, but may be unsustainable long-term.

2030–2035: Unclear Business Model (<$100B potential)

Even if someone builds the best robotic brain, monetisation is hard because it sits far from the end-user value capture — much like PC and smartphone OS platforms that were built in-house (Windows, iOS), absorbed into ecosystems (Android & Google), or open-sourced (Linux). Robotics is even tougher: many OEMs are trying to build internal AI models, and unlike LLMs trained on internet data, robotic AI needs contact-rich physical-world feedback that’s easier to collect with owned hardware. Monetisation paths could be similar to AV players — hardware + software (Mobileye, Horizon Robotics), going downstream into applications (e.g. Palladyne AI going into defense, or Waymo and Momenta going into ride-hailing), or tooling/integrations (fleet management; Red Hat & Linux analog) — but most of these break the “software-only” categorisation. A cleaner software-only wedge is dev tools for robotics, especially data management, similar to Scale.ai or Surge.ai but tailored to robotics.

After 2035: Focus on Dev Tools ($100B+ potential)

As robots mature, robotic AI models likely commoditise (similar to LLMs today), pushing pure “robotic brain” developers toward applications or full-stack solutions. At the same time, a much larger robotics industry will create durable demand for software infrastructure — data management, security, compliance, and adjacent tooling — where $100B+ valuation outcomes become plausible.

General-Purpose Robotic Arms

Robotic arms are the most basic, proven embodiment of practical robots, and they’ve been a backbone of industrial automation since the dawn of the industry, dominated by established players like FANUC, KUKA, ABB, and Yaskawa. But traditional arms are also increasingly commoditised: they require customised pre-programming and integration work, which constrains scalability across sites and use cases.

By 2030: Generalist Approach (~$10B potential)

A new wave of VC-backed unicorns like Flexiv, Agile Robots and Aubo Robotics is targeting the scalability bottleneck by pairing arms with more advanced foundation/VLA models that can follow instructions and generalise across objects, scenes, and modest task variation. Combined with tactile sensing and force control, dexterous hands, and collaborative operation, these players aim to expand the feasible task set and take share from legacy manufacturers. In parallel, software layers and no-code platforms reduce integration and programming friction, making deployments faster and more repeatable.

2030–2035: Stationary Assistants (~$50B potential)

As robot adoption accelerates, the TAM for arms expands, splitting between cheap traditional arms and more flexible “general-purpose” arms. Advanced service-robot and humanoid developers will likely keep using traditional arms (or build their own), while general-purpose arms are adopted broadly by industrial incumbents and less-automated businesses that want robotics without full mobile platforms — restaurants, warehouses, pharmacies, etc. These systems effectively become stationary assistants: cheaper than mobile service robots/humanoids, scalable across many small workflows, and defensible through a developer-friendly platform plus accumulated datasets for continuous model improvement. The margin risk is that many deployments will be owned by task-specific robot operators with their own capabilities, where the arm becomes “middleware,” compressing hardware and platform margins.

After 2035: Infrastructure Play ($100B+ potential)

General-purpose arms likely become an industry standard with a few dominant vendors — similar to ABB/KUKA/FANUC today, but with stronger software-enabled moats and better scalability. Even so, because these platforms sit relatively far from end-customer value capture — and their lock-in is unlikely to match NVIDIA/CUDA-level control — they may struggle to scale far beyond ~$100B, closer to the ceiling implied by today’s large manufacturing platforms (similar to Foxconn).

Laboratory Automation

The market seems niche, but it’s large enough with some incumbents reaching significant scale, e.g. Thermo Fisher Scientific ($220B+ market cap), Beckman Coulter (reached $5.8B in sales at some point), Roche Diagnostics ($14B+ revenue). This segment includes discovery/R&D labs (tech R&D, research institutes, big pharma) and testing / quality control labs across factories.

By 2030: New Disruptors (+$10B potential)

Demand is driven by large share of manual processes (up to 90% for some cases) and high cost of lab personnel. Scalability is decent for horizontal “AI-powered lab” stacks (e.g. Periodic Labs, OpenTrons, MegaRobo) but slowed by heterogeneous lab layouts, labware variation, and integration into existing instruments / management systems (LIMS), while task-specific systems like Equashield can scale faster within a narrow workflow.

2030–2035: Fully Autonomous Labs (~$100B potential)

The mass adoption of “fully autonomous labs” with a closed-loop from experiment design to execution, measurement and learning, challengers will be gradually replacing incumbents by accelerating entire R&D and quality control cycles, along with providing modular robotic blocks with higher scalability. However, this process will take more than a decade, because the existing systems are sticky because of the existing experimental datasets and regulated workflows.

After 2035: Mature Niche ($200B+ potential)

As autonomous lab operations become the industry standard, new disruptors could reach — and potentially surpass — incumbent scale ($200–300B market caps). Even if the overall TAM isn’t massive, the combination of high switching costs and the strategic value of owning the full R&D and quality control loop should let the leading platforms capture an outsized share of industry profit pools.

This table summarises the estimates above. I believe every segment listed here is “VC-scalable”, but the valuation upside varies meaningfully by segment and time horizon.

For early-stage investments in startups priced below $50M, even after ~90% dilution until a liquidity event, a ~20x return over the next 5–7 years and up to ~200x over 10–12 years is still a reasonable base-case scenario.

For growth-stage funds targeting a “safe” 10x in 4–5 years and assuming ~50% dilution (i.e., 2–3 new rounds before liquidity), underwriting becomes much more sensitive to entry price:

$500M entry: almost any segment can still deliver 10x.

$1B entry: 10x remains plausible for autonomous driving, aerial & marine systems, medical robots, warehouse automation, and robotics software.

$5B entry: 10x is realistically only possible in autonomous driving or aerial & marine systems.

These are best-case outcomes for the category leaders and don’t account for downside risks (execution, regulation, geopolitics). They also exclude more bullish, hype-driven scenarios with inflated valuations — which matter in practice, but are hard to predict.

Final Thoughts on China

Some time ago, I published an article on how China nurtures its global champions in areas like solar panels, wind energy and electric vehicles. In that piece, I’ve highlighted robotics as the sector which has the highest potential to give birth to the “next BYD, Bytedance and renewables giants”.

China currently concentrates the largest share of the global robotics hardware supply chain within its borders — often even within single industrial clusters like Shenzhen and the Yangtze River Delta.

As robotics and autonomy become the next major technology wave reshaping every industry, China’s role in that global transition becomes even more central — bringing both outsized opportunities and structural challenges.

Here’s how I see China’s role in global robotics playing out over the coming decades:

• Chinese hardware OEMs will remain the backbone of the global supply chain — from basic components to robotic arms and increasingly sophisticated service robots and humanoids. Western suppliers will still matter in sensitive/strategic domains, but China will dominate mass-market use cases.

• Industry-specific deployments will be led by local companies — integrators and robotics-powered operators, largely sourcing hardware from Chinese (and other) OEMs while building workflow, distribution, and service layers locally.

• China will be strong in consumer-facing applications (robotaxis, home humanoids) with major players successfully expanding abroad. In the near-to-mid term, Europe is likely the primary overseas market as the largest mature region while the US becomes more sensitive to Chinese suppliers; longer term, gravity may shift toward emerging markets (APAC, LATAM) as they mature and as Europe becomes more self-reliant and more skeptical toward China.

• West will lead in premium, regulated niches (medical robotics, laboratory automation), with Chinese players becoming more competitive but primarily focused on domestic demand plus selective emerging markets.

• China will catch up in robotics AI with its open-source approach. Robotics AI models are the main bottleneck and are currently dominated by US companies, but China has an advantage in physical-world datasets that matter for training embodied systems. As with LLMs, China may produce strong competing robotics models, yet many will likely be open-sourced with limited direct monetisation and ultimately absorbed into larger platforms.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Methodology:

Companies from the following segments are included in the research:

• autonomous vehicles, drones and other unmanned mobility platforms;

• general-purpose robots (robotic arms, robodogs, humanoids, rovers etc) and special-purpose robots (medical robots, warehouse robots etc);

• developers of software or hardware used specifically for robots;

• only companies with robotics as their core products are included;

• if a company is engaged in several segments, it is assigned to a category of its core product;

• non-VC-backed companies (like KUKA, FANUC etc) are excluded;

• VC-backed divisions or spin-outs of large corporations are included (e.g. Waymo / Google, Mobileye / Intel, Yinwang / Huawei), but non-VC-backed divisions are excluded (Optimus / Tesla).

Information on the valuations of private unicorns is retrieved from press releases or VC databases, such as CBInsights, Crunchbase and ITjuzi.

Sources for TAM estimations:

• Humanoids & Animal Robots. ABI Research — Humanoid Robot Market Size, 2024 to 2030

• Humanoids & Animal Robots. Morgan Stanley — Humanoid Robot Market Expected to Reach $5 Trillion by 2050

• Humanoids & Animal Robots. MarketsandMarkets — Humanoid Robot Market Size, Share, Industry Report Trends, 2025 To 2030

• Unmanned Aerial & Marine. Spherical Insights — Top 50 Commercial Drone Market Size, Statistics Report Till 2035

• Warehouse Automation. Data Bridge Market Research — Warehouse Robotics Market Size, Share, and Trends Analysis 2032

• Warehouse Automation. IDTechEx — Mobile Robotics in Logistics, Warehousing and Delivery 2024–2044

• Warehouse Automation. IDTechEx — Mobile Robots, Autonomous Vehicles, and Drones in Logistics, Warehousing, and Delivery 2020–2040

• Warehouse Automation. LogisticsIQ — Warehouse Automation Market

• Medical Robotics. Data Bridge Market Research — Medical Robots Market Size, Trends, Growth Report 2032

• Medical Robotics. Fortune Business Insights — Medical Robots Market Size, Share & Forecast Report, 2034

• Hardware Components. WiseGuyReports — Robot Parts Market

• Service Robots. Fortune Business Insights — Service Robotics Market

• Service Robots. Mordor Intelligence — Service Robotics Market

• Robotic Arms. Data Bridge Market Research — Robotic Arm Market

• Software for Robots. Market Research Future — Robot Software Market

• Software for Robots. Mordor Intelligence — Robot Software Market

• Laboratory Automation. Vision Research Reports — Laboratory Robotics Market

• Laboratory Automation. Grand View Research — Lab Automation Market

Across both developed markets such as Europe and emerging regions like Southeast Asia, the traditional VC model is under growing strain. Constrained exit markets and limited liquidity have made LPs increasingly reluctant to commit capital to another ten-year cycle with local early-stage VC funds. As a result, both LPs and GPs are reallocating toward more established asset classes, including mid-market buyouts, traditional private equity, and private credit.

At scale, only two ecosystems still offer the full venture playbook: the United States and China. While global investors remain deeply familiar with the US market, interest in China’s VC ecosystem has only begun to meaningfully recover over the past year.

2025 was phenomenal for China.

The year began with Chinese startup DeepSeek releasing its R1 model, which surprised the global tech community with its exceptional efficiency and performance — on par with leading US AI labs. As I noted in a previous article, while DeepSeek itself was not a VC-backed story, it created a powerful global reference point — the so-called “DeepSeek moment” — that materially strengthened confidence in China’s deep-tech capabilities among customers, investors, and the media.

That shift in perception helped propel a broader wave of Chinese AI companies, including MiniMax, Moonshot, and Manus AI, enabling them to gain international traction, raise larger funding rounds, pursue public listings, and, in some cases, attract acquisition interest from major US technology firms.

After 2022, despite COVID lockdowns and a prolonged regulatory reset in the tech sector, China’s venture activity continued to expand. By 2024, the number of investment rounds in China had reached parity with the entire European region, including the UK and Israel.

By Q4 2025, China’s tech deal volume exceeded Europe’s by roughly 80% and had nearly caught up with the US.

China also remains by far the largest VC/PE ecosystem in APAC, surpassing India — the region’s second-largest market — by approximately six times in deal count.

In terms of total capital raised by tech startups, China is broadly comparable to Europe but remains significantly smaller than the US, which attracts more than 3.5× the funding.

Despite raising a similar amount of capital, Chinese startups generated roughly 4x more PE/VC exits than their European counterparts. This was driven primarily by a wave of large public listings, led by the CATL IPO in May 2025 ($134B), alongside other sizeable offerings such as Seres ($26B) and SF Express ($19B).

In total, China accounted for 86% of all tech PE/VC exit value in APAC in 2025.

Focusing only on the top five deep tech verticals — AI, compute, robotics, energy, and space — Chinese startups even surpassed their US counterparts in 2025 in terms of PE/VC exit value. Beyond the landmark CATL IPO, the growth was driven by a combination of large primary listings, including Moore Threads ($7.5B) and MetaX ($5.9B), as well as secondary listings in Hong Kong such as WeRide ($6.8B) and Pony.ai ($6.0B).

China contributed to 94% of all deep tech PE/VC exit volumes in APAC in 2025.

Historically, China has recorded fewer exits by deal count than Europe or the US, largely because IPOs — rather than M&A — have been the dominant liquidity path for Chinese startups. While IPOs are less frequent, they tend to be significantly larger in size than typical M&A transactions.

In 2025, this dynamic shifted meaningfully. China nearly reached parity with Europe in total PE/VC exit count and surpassed Europe — while almost matching the US — in the number of deep-tech exits. This acceleration was driven by a sharp increase in both IPO activity (up 70% year-on-year) and M&A transactions (up 36% year-on-year).

Despite the early momentum of Chinese AI startups, the US remains by far the largest VC market for AI — both in funding volume and deal count. This dominance is driven by a concentration of mega-rounds, including OpenAI ($40B and $7.6B), xAI ($5.3B), and Databricks ($5B), as well as a much broader base of VC-backed AI application startups than in China.

AI adoption in China is also robust, but it’s not as broadly productised through VC-backed startups as in the US. Instead, it’s mostly concentrated within large corporations or even state-owned companies.

Even so, by 2025 Chinese AI startups surpassed their European peers in deal count. Within APAC, China accounted for approximately 62% of all AI investment rounds, compared with around 20% for India and 12% for Japan.

As with AI, the US remains the largest source of capital for next-generation compute technologies, driven primarily by a handful of mega-rounds. These include large financings for emerging compute infrastructure and accelerator players such as Crusoe ($11.6B), Lambda ($2B), Cerebras ($1.1B), and PsiQuantum ($1.0B).

China, meanwhile, has led global compute investment by deal count over the past several years. However, the majority of Chinese startups operate across the traditional semiconductor value chain — covering wafer manufacturing, testing and packaging equipment, EDA tools, and other critical components of the domestic chipmaking ecosystem.

There are also more advanced players, e.g. the next generation of Chinese GPU providers like MetaX, Moore Threads, Biren, Cambricon, Enflame and others that are building Chinese alternatives to NVIDIA.

In robotics and autonomous driving, China has already emerged as a formidable force — both in the number of startups raising capital and in total funding volume. Although US has massive companies like Waymo (in talks to raise $15B at $100B+ valuation) and Figure (raised at $39B valuation), China had 50% more robotics investment rounds than the US in 2024 and 3.5x more rounds in 2025.

These investments span a wide range of segments, including autonomous vehicles, humanoid robots, drones, warehouse automation, robotic arms, and core hardware components across the robotics supply chain.

Within APAC, China accounted for approximately 87% of all robotics investment rounds, with the remaining activity fragmented across South Korea, Japan, and India, each contributing roughly 3–4%.

In 2025, India reached 60% of China’s VC/PE investment volumes in energy and climate startups, with projects like Greenko ($1.4B), Inox Clean Energy ($550M), Exide Energy ($420M).

Meanwhile, US is leading with mega projects like Sempra ($10B), Origis Energy ($1.3B), Commonwealth Fusion ($860M), X-Energy ($720M).

While US space- tech startups attracted more capital than their Chinese counterparts in 2025 ($4.3B vs. $2.4B), China overtook the US in deal activity (187 vs 139) with the rounds like Space Pioneer ($350M), Galactic Energy ($350M), GeeSpace ($280M), Minospace ($147M) and iSpace ($140M).

India is also becoming a stronger space tech ecosystem with 41 deals in 2024, compared to 73 deals in Europe. Largest deals included Raphe mPhibr ($100M), RMSI ($56M), Digantara ($50M) and Rangsons Aerospace ($34M).

Despite the improving dynamics of China’s VC and deep tech ecosystem, the share of investment rounds involving overseas capital declined sharply — from 27% in 2021 to around 11% in 2025. This contraction has been driven primarily by geopolitical constraints and a persistent information gap around the Chinese startup landscape.

Importantly, the 11% figure for 2025 also includes overseas funds managed by Chinese GPs — such as GGV and GSR — which only partially qualify as cross-border investors.

The broader trend is a growing reluctance among foreign funds to invest directly into Chinese startups from their flagship global vehicles. This analysis does not capture overseas LP commitments to China-focused GPs, which face fewer formal restrictions but remain largely opaque due to limited public disclosure.

Most Western funds — primarily US-based — continue to concentrate their APAC deep tech exposure in India and Japan. These markets are perceived as carrying lower geopolitical risk, despite being less scalable than China.

Overall, Western participation in Asian deep tech remains limited. In Q4 2025, the most active US-based investor in Asia, Accel, completed no more than three Asian deep-tech investments. By contrast, Singapore-based Antler—the most active Asian investor in the US and Europe — backed 11 Western startups over the same period, while Peak XV and Mubadala each made six investments.

The US remains the primary destination for Asian deep tech investors, led by Singapore-based funds such as Antler, Temasek, GIC, and Xora. This marks a sharp reversal from 2020–2021, when Chinese technology groups — including Tencent, Alibaba, and Baidu — were among the most active Asian backers of Western deep tech startups. This comparison does not capture Chinese GPs that have established onshore operations directly in the US.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn.

Let’s connect and explore how we can improve together.

• China’s PE/VC activity has growth the fourth quarter in a row, nearing the US deal count (2,376 vs 2,588), while the funding gap remains wide ($71B vs $20B).

• US AI funding keeps plummeting from $47B in Q1 to $16B in Q3; while the number of AI rounds is steadily growing both in the US and China.

• China is significantly outpacing Europe and the US by the number of investment rounds in robotics and compute technologies (3-5x).

• China is also leading in space tech for the 4th quarter in a row (by deal count).

• Western investors are still cautious about investing directly in Chinese startups, doubling down on markets in APAC like India, Japan and Australia.

• While Hong Kong IPOs went up and overall VC/PE exits in Asia are slowly recovering, they still trail the US and remain roughly on par with Europe.

US was by far the world’s largest tech ecosystem in terms of PE/VC funding amount in Q3 2025 with major investment rounds like the $5B+ raise of Elon Musk’s xAI. Tech PE/VC funding in Europe also performed strongly, significantly outpacing China the first time for the past three years. As for the deal count, China is rapidly catching up with the US and by far surpassing Europe.

US has been fuelling the global AI investments for the past two years, however, the AI funding in American startups is dropping the second quarter in a row.

Is the AI bubble finally going to burst soon?

We’ll wait and see. But the number of AI startups securing funding — both in the US and China — keeps growing, which gives hope that the real AI adoption is closer than we think.

Some of the interesting AI rounds from Asia below:

• 🇨🇳 $273 round from Chinese state funds: Minimax (稀宇科技), developer of general large models of different modalities, including a trillion-parameter MoE text large model, a speech large model, and an image large model.

• 🇨🇳 $140M round from Pudong, Zhangjiang Group: Zhipu (智谱), China’s leading developer of the full-stack open-source LLM.

• 🇨🇳 $111M seed round from SAIF, Linear, Meituan: TARS (它石智航), a company developing large language models for embodied AI, as well as the integration of software and hardware.

• 🇯🇵 $50M Series E from Goldman Sachs: LegalOn Technologies, an AI platform helping lawyers improve corporate legal operations efficiency with advanced legal AI tools.

• 🇰🇷 $45M Series B from KDB: Upstage, an AI-driven language models and document processing engines that streamline data extraction, analysis, and workflow management.

Compute funding in the US also started to drop, which I think is connected with the falling AI investments. However, it is still more than 3x larger than in Europe and China, with such major round as Vantage Data Centers ($1.6B), Cerebras ($1.1B) and PsiQuantum ($1B).

Meanwhile, China is far ahead in terms of the number of compute-related investment rounds — 4-5x larger than the US and Europe. It’s mostly driven by the local government push to build a fully self-sufficient supply chain for chipmaking.

A few standout compute investments in Asia include:

• 🇰🇷 $250M Series C from Arm Holdings, Samsung: Rebellions, a provider of AI accelerator chips and data center systems designed for efficient inference in large‑scale applications.

• 🇨🇳 $210M Series C from Tencent, Baidu: Xizhi Technology (曦智科技), a provider of hybrid optoelectronic computing power, offering a series of computing power transition solutions, starting from the three core technologies of photon matrix computing (oMAC), on-chip optical network (oNOC), and inter-chip optical network (oNET).

• 🇰🇷 $123M Series C: FuriosaAI, a developer of data center accelerators for the most advanced AI models and applications.

• 🇨🇳 $42M Series C from Legend Capital: SJI Semi (尚积半导体), developer of a new semiconductor front-end equipment.

• 🇨🇳 $42M Series A from Primavera Capital: PSR Semi (研微半导体), a startup specializing in high-end thin film deposition equipment, with product lines covering thermal ALD, PEALD, silicon epitaxial, silicon carbide epitaxial, and PECVD equipment, etc.

China continues to dominate global robotics funding, outpacing all regions despite the $1B Series C of US humanoid leader Figure AI. In Q3 2025, Chinese robotics startups closed 3x more deals than their US peers and 5x more than Europe.

Some of the recent robotics investments in China and the of Asia include:

• 🇨🇳 $140M Series A from JD, T-Capital, Stone Venture: Engine AI (众擎机器人), a general-purpose humanoid robot startup with the full-stack development of all the core components of the body to embodied intelligence and operation control algorithms.

• 🇨🇳 $140M Series A+ from Hongshan, Legend, Meituan, Alibaba: X Square (自变量机器人), a startup building general intelligent agents that can operate precisely based on large robot models.

• 🇨🇳 $91M Series A+ from IDG, Cathay, Meituan, MiHoYo: Galaxea AI (星海图), an AI startup developing embodied intelligence basic models and embodied intelligence robots.

• 🇨🇳 $70M Series C+: Deep Robotics (云深处科技), developing foot type robots, including quadruped bionic robots that can run fast and the series of robots for industrial applications such as power stations, factories, pipe gallery inspections, emergency rescue, and fire reconnaissance.

• 🇯🇵 $57M Series D from MUFG Bank: SkyDrive, an automotive start-up that develops flying cars and logistics drones.

Q3 2025 saw a dramatic increase in climate funding in the US and Europe, mostly driven by two large infrastructure deals: TenneT in the Netherlands ($11B) and Sempra in the US ($10B).

Q3 was also the first quarter for a long time when China outpaced Europe and the US by the number of climate investment rounds.

A few notable climate deals in Asia include:

• 🇮🇳 $405M PE round from Green Climate Fund: Vertelo, offers complete fleet electrification services.

• 🇦🇺 $329M PE round from KKR: CleanPeak Energy, clean energy solutions for commercial and industrial sectors, focusing on sustainability.

• 🇸🇬 $72M round: I-Pulse, a high-pulsed-power technology company that develops innovative commercial applications for pulsed power technologies.

• 🇨🇳 $70M seed round from Legend Capital, Alibaba: Nova Fusion (诺瓦聚变) develops small, distributed, and modular nuclear fusion devices, adopting adopt the advanced field reaction (FRC) magnetic compression technology route to develop small modular reactors (SMR).

• 🇨🇳 $42M Series A from Qiming: Sany Hydrogen (三一氢能), developer of water electrolysis equipment for producing hydrogen, focusing on downstream customers such as large-scale wind-solar coupled hydrogen production, chemical engineering, metallurgy, transportation, and hydrogen blending in natural gas.

The US-China space race is intensifying. Both countries have comparable amount of PE/VC funding for space tech startups for the past two years. At the same time, China keeps leading in terms of the number of space tech rounds the fourth quarter in a row.

It’s worth noting that India’s number of space tech deals is growing steadily and now only ~40% smaller than in Europe.

Here are some standout space tech deals in APAC from Q3 2025:

• 🇨🇳 $336M Series D: Galactic Energy (星河动力), developing low-cost commercial rockets with a possible payload of 350 kg, focused on bringing microsatellites to the low-earth orbit.

• 🇨🇳 $280M round from a state EV-focused fund: Geespace (时空道宇), a global low-orbit communication constellation of 72 satellites serving real-time data communication for mass consumers and various industry applications.

• 🇯🇵 $10M Series C: Pale Blue, a developer of water propulsion systems and satellite mobility infrastructure for the aerospace industry.

• 🇮🇳 $10M Seed round: Astrobase, India’s first reusable, medium-lift orbital launch vehicle—purpose-built for on-demand launches, constellation deployments, sovereign missions, and global satellite infrastructure.

• 🇨🇳 $3M Pre-A round from Vertex: Dragonfly Wing (蜻蜓翼行), a tilting wing eVTOL, the first technical team in China to complete the development of 30kG-class tiltrotor flight control technology.

India is the most popular investment destination in Asia for Western deep tech investors, along with Japan and Australia.

US is the most popular destination for Asian deep tech investors outside of their home region. Indian and Singaporean allocators were the most active in Q3 2025.

Worth noting that none of the Chinese funds got into the TOP 10 due to geopolitical constrains. It’s reasonable to assume that some transactions with Chinese investors are also not reported for the same reason.

The US remains the largest exit market for VC/PE investors, leading both in the number of exits and total exit value. In Europe, liquidity is gradually recovering after a prolonged slowdown.

China’s exit value saw a temporary correction in Q3 following CATL’s massive IPO in Q2, though the overall trend remains upward.

2025 has been a strong year for Hong Kong’s capital markets — IPO volumes surged by over 200% in Q3, according to KPMG. However, the number of VC/PE-backed Chinese IPOs declined from 44 in Q2 to 29 in Q3.

At the same time, PE/VC-backed M&A activity in China reached a multi-year high, with 92 transactions in Q3 compared to the usual 40–60 range in prior quarters — a clear sign of growing market consolidation.

Below are some of the Q3 exit highlights from APAC:

• 🇨🇳 IPO at a $2.4B market cap: Geek+ (极智嘉), leading developer of robotics solutions for warehouse logistics. Exit for Intel, B Capital, Vertex, Volcanics, Gaorong, Vertex and others.

• 🇨🇳 IPO at a $3.5B market cap: BEST (屹唐股份), manufacturer of wafer processing equipment used for dry de-gumming, dry etching, rapid heat treatment, millisecond-level annealing and other solutions. Exit for Hongshan (ex-Sequoia China), IDG, Shenzhen Capital.

• 🇨🇳 IPO at a $1.5B market cap: Fortier Tech (峰岹科技), a startup specializing in developing chips for motor drivers used for industrial equipment, motion control, power tools, consumer electronics, intelligent robots, and other drive control fields. Exit for Xiaomi, Legend Capital, Shenzhen Capital.

• 🇨🇳 $645M acquisition: PhiGent Robotics (鉴智机器人), a provider of next-generation autonomous driving systems focused on visual 3D understanding, got acquired by a Chinese public mobility tech company NavInfo ($3B+ market cap).

• 🇯🇵 IPO at a $160M market cap: Axelspace Holdings, a pioneer in microsatellite technology. Exit for SBI, MUFJ, SMBC, Keio Innovation Initiative, ITOCHU.

• 🇰🇷 IPO at a $150M market cap: AUTOCRYPT, a developer of end-to-end mobility security solutions for autonomous vehicles. Exit for KB Investment, Ulmus, Shinhan, Pathfinder H, BSK.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn.

Let’s connect and explore how we can improve together.

Check out the full version here.

In my previous article, I argued that DeepSeek’s sudden rise turned out to be less revolutionary on the global scale than the media suggested. Rather, it served as a catalyst for trends already in motion — declining LLM prices, the expansion of open source, and technical innovations such as Mixture-of-Experts (MoE) and Multi-Head Latency Attention that made its models highly efficient.

Yet, the disruption DeepSeek triggered within China’s AI ecosystem — and its pivotal role in shaping both domestic and global trajectories — proved far more significant.

Red Ocean of Chinese LLMs

China is home to over 200 companies developing their own LLMs, including major tech giants such as Alibaba, Tencent, Baidu, and Bytedance. Alongside them, a new wave of next-generation AI startups—often referred to as the “Six Tigers”—has emerged with a singular focus on advancing frontier LLMs. These include Zhipu AI, Moonshot AI, Baichuan, Stepfun, Minimax and 01.ai.

The first major shock came on May 6, 2024, with the release of DeepSeek-V2. Its API launched at just $0.14 per 1 million input tokens—a dramatic 5–10x lower than the prevailing market rates at the time.

This immediately triggered a price war. Alibaba quickly slashed the price of its Qwen model from $1.10 to $0.07 per million tokens, and just a week later, Bytedance entered the fray, launching its new model, Doubao, with an aggressively low entry price of $0.04 for the lite version.

By the time DeepSeek’s R1 and V3 models gained global attention in December 2024 and January 2025, the average price of LLMs in China had already fallen by 92% compared to May 2024. Following the R1 release, Chinese providers were forced to cut prices further—though the drop in early 2025 was less dramatic, averaging around 30%.

The response to DeepSeek’s aggressive pricing was far more intense in China than in the rest of the world — largely because, unlike OpenAI, xAI, or Anthropic, domestic AI companies viewed DeepSeek as a direct competitor. As I explored in my recent piece on China’s strategy for nurturing global champions, price wars are a familiar pattern across most Chinese industries — from solar panels and EVs to consumer electronics.

By January 2025, DeepSeek R1 had become the top-performing LLM in China, but rivals like Moonshot AI, Stepfun, and Baichuan were not far behind, creating a fiercely competitive domestic landscape.

Since January 2025, only Alibaba and Minimax have released new models that rival DeepSeek R1 in performance.

Survival of the Fittest

If 2024 was the year of AI price wars in China, then 2025 became the year of strategic pivots. Beyond DeepSeek, Alibaba, and Minimax, only a handful of players — such as Bytedance, StepFun, and Moonshot AI — chose to stay in the game of developing large, high-performance models, while many others began to scale back or refocus their efforts.

Zhipu AI is focusing on commercial traction through developing enterprise and consumer grade applications ahead of its upcoming IPO. Baidu and Tencent doubled down on integrations with DeepSeek and leveraging its extensive ecosystem. 01.ai started focusing on enterprise-grade solutions powered by DeepSeek, while Baichuan prioritized healthcare applications. Minimax, despite still heavily investing in its foundational model, leverages its video and image generating consumer applications.

I think all these pivots make a lot of sense. DeepSeek positions itself as a research-heavy infrastructure company, rather than a B2C customer facing app. And it clearly can’t do both, considering that the resources of its parent company are not as vast as OpenAI or Bytedance. That is why despite DeepSeek’s growing B2B adoption, its consumer web traffic is decreasing after riding the hype wave.

This level of specialization among Chinese LLM players stands in stark contrast to the global landscape, where leading providers like OpenAI, Anthropic, and xAI typically pursue end-to-end strategies—building both core infrastructure and industry-specific applications.

Open Source is the New Normal

Another consequence of DeepSeek’s rise that I mentioned above is that it pushed many previously closed-source LLMs in China to adopt partially or fully open-source strategies.

On one hand, DeepSeek has set a new industry benchmark by offering basic LLM infrastructure for free. On the other, intensifying competition in China is likely to push domestic LLM providers to expand globally. To succeed internationally, many of them will need to embrace open-source strategies, as discussed earlier.

In parallel, DeepSeek appears to have improved the overall perception of Chinese AI models—even outside China. For instance, I've observed a growing number of European startups adopting Alibaba’s Qwen, signaling a shift in sentiment and trust.

What Comes After DeepSeek

DeepSeek’s impact on Chinese companies building AI applications has also been profound — both directly and indirectly.

Lower infrastructure costs

The most immediate effect has been the LLM price war triggered by DeepSeek’s ultra-low pricing. Unlike their global counterparts, Chinese developers are not limited to open-source versions; they can leverage cloud-based LLMs at a fraction of the cost—a major advantage, especially given the complexities of self-hosting under current U.S. export restrictions.

This creates a unique cost advantage for Chinese AI developers, allowing them to build and scale more efficiently. However, it may also discourage international expansion, since low-cost LLM access is largely restricted to domestic users—or more precisely, to customers comfortable with Chinese cloud infrastructure handling their data.

Better funding environment

Over the past four years, Chinese AI startups have faced a tough funding environment, with venture and private equity investment in sharp decline.

AI funding in China began to show signs of recovery in 2024—likely aided by falling infrastructure costs. However, the real momentum shift came in the first half of 2025, when the number of funding rounds by Chinese AI startups rebounded to levels not seen since early 2022. While the total capital raised hasn't grown significantly, this trend suggests a reallocation toward early-stage startups rather than continued concentration in already well-funded players.

In my view, the "DeepSeek moment" in January 2025 was a key catalyst. Its international breakthrough not only shifted the global narrative on Chinese AI, but also revived investor confidence in the broader ecosystem. Given how closely investor sentiment tracks media perception, DeepSeek’s global spotlight likely encouraged both domestic and international funds—including Temasek, Prosperity7, EQT, and even Benchmark—to re-engage with Chinese AI startups.

Flexible startup culture

DeepSeek is truly unique in its culture that is drastically different from many other tech companies in China like Bytedance or Alibaba. This topic was covered in much detail by JS Tan (here and here) and Jon from Asianometry (here), so I’ve just summarized their points below.

The key takeaway is that DeepSeek’s founder, Liang Wenfeng, is challenging the conventional belief that Chinese companies need high-profile executives, massive teams, and grueling work culture to compete globally.

In many ways, DeepSeek operates more like a U.S. startup than a traditional Chinese one—emphasizing agility, creative problem-solving, and flexible execution over brute force and hierarchy.

Inspiration to go global

Bytedance was arguably the first “truly global” Chinese tech company, with TikTok reaching around 760 million monthly active users (MAU) in China and 1.8 billion MAU internationally. However, Bytedance was founded in 2012, and its global expansion only began in 2017 with the acquisition of Musical.ly. Its first product, the news app Toutiao, was entirely China-focused.

In contrast, DeepSeek has demonstrated that a Chinese company can be global from day one—not just in market reach, but also in technical competitiveness. It has contributed to a more positive global image of China’s tech ecosystem.

DeepSeek’s success with international users—amplified by strong global media coverage—encouraged other leading players like Manus and Minimax to target global markets early. While DeepSeek wasn’t directly responsible for their success, it undoubtedly helped legitimize their global ambitions and accelerated their credibility abroad.

In contrast to the US or Europe, most AI application startups in China that raised money in the first half of 2025 were hardware-related, developing either robots (Flexiv, Noematrix), drones (i-Kingtec, SkySys), autonomous vehicles (Zelos, Boonray, Yihang) or industrial solutions (Ensonic).

A wave of early-stage AI agent startups has emerged in China following DeepSeek’s breakthrough—many of them raising their first institutional rounds in 2025 and targeting global markets from day one. Companies like Shulex, WaveSpeed, and LynxAI are among this new cohort. I’m sure almost nobody outside of China has ever heard about them yet. Some of these companies might fail quickly, but the others may become the “next Manus”.

DeepSeek’s success has redefined the global ambition of Chinese startups, proving that it’s possible to compete internationally without raising billions or relying on government backing. In interviews, Liang Wenfeng has hinted that this very aspiration—to build a lean, globally relevant AI company from China—was one of his core motivations for starting DeepSeek.

Challenges remain

There are still some other aspects that have prevented Chinese software companies from winning the global markets in the past and that DeepSeek is not able to solve.

• Geopolitics and negative perception. Unfortunately, it is what it is. I appreciated how Rita Liao, former TechCrunch reporter, expressed her frustration with Chinese startups hiding their identity—only to note later in her Substack that, post-DeepSeek, more founders are embracing the label of “Chinese challengers.” Still, the skepticism—and in some cases, outright hostility—from Western customers is unlikely to fade without a fundamental shift in geopolitical relations. Until then, even neutral or apolitical companies like DeepSeek will continue to face accusations of “ties to the Chinese military,” regardless of how transparent or open they try to be.

• B2B Sales and go-to-market. ​​One reason Chinese consumer tech has scaled globally is that most users never interact with the companies behind the products. Few Western consumers know that Red Note, Temu, or Shein are in fact Chinese brands. In contrast, B2B sales require trust-building, founder-led conversations, and hands-on customer support—especially in the early stages. This is where cultural and political barriers become real friction points.

Final Remarks – True Meaning of DeepSeek

In my view, DeepSeek’s greatest impact lies not in technological novelty, but in its symbolic value for China’s tech ecosystem. A small team based in Hangzhou—not Silicon Valley—composed of recent graduates from Chinese universities—not Stanford or MIT—and with no prior experience at Meta, Google, or Amazon, managed to build an LLM that rivals top-tier global models while using a fraction of the resources.

DeepSeek didn’t invent the Mixture of Experts (MoE) architecture—that dates back to the 1990s. Nor was it the first to apply MoE to AI; Google Brain, DeepMind, and later Mistral led that adoption in the early 2020s. But DeepSeek did what the best companies do: they took an existing idea and made it significantly more efficient and scalable.

In that sense, the "DeepSeek phenomenon" should come as no surprise to anyone familiar with China’s industrial playbook—as I’ve described here. It mirrors the trajectory of Jinko Solar in photovoltaics, CATL in batteries, BYD in EVs, and Bytedance in recommendation algorithms.

DeepSeek and its founder Liang Wenfeng hold symbolic importance for China. After four difficult years for China’s tech sector (2021–2024), they’ve served as a powerful reminder that China remains the world’s second-largest tech ecosystem—with untapped global potential.

While DeepSeek hasn’t disrupted the global AI landscape or directly accelerated the growth of other Chinese companies, its deeper contribution lies in something harder to measure—but far more meaningful: conviction.

• Conviction among Chinese entrepreneurs to build globally competitive companies—without leaving China.

• Conviction among domestic investors to back homegrown talent with global ambition.

• Conviction among international investors to take a fresh look at China—not as a follower, but as a creator of breakthrough technologies.

And that, arguably, is exactly what Liang Wenfeng set out to achieve.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

The article is the first part of the deep-dive series “Was DeepSeek Such a Big Deal?” originally published in newsletter in collaboration with .

Check out the full version here.

Also big thanks to my Runa Capital’s colleagues and Maxime Corbani for sharing valuable ideas for this topic!

In January 2025, an unknown Chinese startup, DeepSeek, stunned the global tech community with the release of its R1 reasoning model—delivering breakthrough performance at a fraction of the typical training cost. This led to a record NVIDIA stock sell-off. Leading media outlets erupted with apocalyptic headlines predicting the downfall of OpenAI and the collapse of America's tech dominance.

Fast forward to early September 2025:

• NVIDIA’s stock is up 22% from its “pre-DeepSeek: levels.

• OpenAI not only survived – it raised another $40B in April 2025, marking the largest private tech funding round in history.

• Other American AI players like xAI and Anthropic continue to attract billions in fresh capital too.

• The launch of DeepSeek’s new R2 model has been repeatedly delayed — originally planned for May, it still has no confirmed release date.

So, was DeepSeek truly as revolutionary as the media portrayed it seven months ago — or was it simply a hype-fueled moment of panic?

In this article, we examine the real long-term consequences of DeepSeek for AI infrastructure and applications globally—separating the media frenzy from DeepSeek’s actual impact.

DeepSeek vs Global AI

When DeepSeek R1 was released, there was an expectation that major LLM companies like OpenAI, xAI, Anthropic and others would re-consider their existing strategies, considering that companies like DeepSeek can train AI models as good as OpenAI, but at a fraction of the cost.

Moreover, DeepSeek’s models are fully open-source, so now anyone can adjust their architectural innovation and training techniques to make their models as efficient and powerful as DeepSeek.

However, Anthropic expects to burn ~$3B in 2025, xAI – $13B and OpenAI more than $14B.

It was widely expected that DeepSeek would trigger a price war in the global LLM market—after all, its models launched at 50–100x lower prices than those of OpenAI, xAI, and Anthropic. But are these companies burning billions just because they were forced to slash prices?

As the chart below shows, LLM prices have indeed declined since January 2025, but this trend began before DeepSeek at the beginning 2024. DeepSeek just accelerated it.

Even after these reductions, the leading global LLM providers still charge 14–15x more than DeepSeek for its APIs (see table below).

Why hasn’t DeepSeek disrupted the business model of global LLM providers?

1. Product development cost. DeepSeek remains laser-focused on research, raw model performance, and the long-term pursuit of AGI—rather than building tailored products for specific industries or use cases. In contrast, companies like OpenAI and xAI are pursuing a dual strategy: pushing the frontier of model development while also investing heavily in security features, multimodal capabilities, user experience, ecosystem integrations, and industry-specific applications in fields like healthcare, finance, and law.

2. Innovators vs followers. DeepSeek and other Chinese LLMs are still in catch-up mode when it comes to matching OpenAI’s performance. However, the path to achieving comparable results more efficiently—particularly through techniques like model distillation—demands significantly fewer resources than pushing the boundaries of frontier research.

3. Extra deployment costs. Most global companies are expected to adopt the open-source version of DeepSeek, rather than its cloud-hosted offering on Chinese infrastructure—primarily due to concerns around data privacy and regulatory compliance.This raises a key question: Is the total cost of deploying and running an open-source model truly lower than that of using closed-source alternatives—especially at scale, with growing inference volumes? Several recent analyses (e.g. here) explore this trade-off in greater detail.

4. Geopolitics and privacy. DeepSeek’s global market share remains small—currently under 2%. As noted earlier, its cloud-hosted version is generally not considered a viable option for international customers due to privacy and data sovereignty concerns. Even when self-hosting the open-source version is technically and economically feasible, many Western companies remain hesitant, largely due to the perception risks of using a Chinese-developed AI model.

When I surveyed European startups for a comment to Sifted.eu, several asked to remain anonymous—not because of technical limitations, but because they feared potential backlash from customers, despite running the model entirely on local infrastructure.

Impact on open source

In my view, DeepSeek’s biggest long-term impact will be on the global open‑source LLM ecosystem.

On one hand, its rise has pushed previously closed players toward open source. We’ve already seen this in China, where giants like Alibaba and Baidu opened parts of their models, and globally with OpenAI releasing its first open-source model, gpt‑oss.

On the other hand, DeepSeek has raised the stakes for existing open-source developers like Meta’s LLaMA and Mistral. Competition is heating up fast. As Kevin Xu suggested in a recent Substack post, Meta’s next AI model may move away from full open-sourcing, and I tend to agree.

Looking purely at performance, pricing, and speed to market over the past year, it’s hard to see why customers would opt for self-hosted versions of LLaMA or Mistral over DeepSeek or Qwen—unless there’s a strong institutional preference to avoid Chinese technology, as is often the case in sectors like government, defense, or finance.

The growing adoption of DeepSeek among B2B users is also reflected in the rising token consumption on third-party hosting platforms.

DeepSeek vs NVIDIA

Finally, about NVIDIA. Was DeepSeek really going to kill it?

Obviously, not.

As we can see on the chart, NVIDIA’s stock rebounded by 22% (as of August 27th, 2025) compared to its levels before the DeepSeek-R1 model was released.

In fact, NVIDIA’s revenue grew 69% YoY in Q1 2025 despite the restrictions of sales to China.

This surge highlights that even as AI training and inference become more efficient, overall demand for compute continues to rise. This is a classic case of Jevons Paradox—where increased efficiency leads to greater overall consumption. As more companies adopt AI and train increasingly specialized models, the demand for high-performance chips continues to accelerate.

As the graph above shows, global demand for semiconductors continues to grow at a steady 15–20% year-over-year—despite the emergence of more efficient models like DeepSeek.

Interestingly, while China’s share of global semiconductor sales had been on a downward trend, this reversed after the release of DeepSeek-R1 in January 2025. Since then, China’s semiconductor demand has begun climbing again, likely driven by increased local training and deployment of AI models.

At the same time, the US government imposed restrictions on NVIDIA’s chip sales to China, while Chinese authorities discouraged the use of its H20 processors—going as far as barring major firms such as Tencent, Alibaba, and ByteDance from purchasing them over data security concerns. As a result, Chinese companies are increasingly turning to domestic alternatives like Huawei, UNISOC, and Enflame, a shift that is likely to further erode NVIDIA’s market share in China over the long term.

Of course, China’s transition to domestic chips has caused delays in AI development. Some experts even link the postponed release of DeepSeek’s new R2 model to the ongoing shortage of high-end processors in the country.

Nevertheless, DeepSeek’s upgraded V3.1 model released in August 2025 is already compatible with existing Chinese chips and some Chinese automakers managed to replace 100% of their demand for chips with domestic suppliers. This is a promising development for China’s AI industry becoming self-reliant in the near future.

Final Thoughts

Despite the overblown media narratives and panic on the public markets back in January 2025, the actual impact of DeepSeek on the global tech ecosystem was slightly exaggerated:

• OpenAI, Anthropic and others still raise billions and burn them in a few months

• LLM API prices are decreasing, but they’ve been decreasing before DeepSeek

• NVIDIA’s business is flourishing more than ever with revenue growing over 70%

As with most hype cycles, the initial expectations around DeepSeek proved difficult to satisfy, but its longer-term effects are steadily unfolding. DeepSeek has already reshaped the open-source LLM landscape by increasing pressure on incumbents such as Meta and Mistral, while nudging closed-source players like OpenAI and Alibaba toward greater openness.

While its direct impact is not always immediately visible, the influence of the “DeepSeek phenomenon” continues to accelerate global AI adoption, pushing LLMs toward becoming a true commodity—much like electricity or internet connectivity in earlier eras.

Even more important is DeepSeek’s profound impact on China’s AI and broader tech ecosystem, and the implications this holds for its position on the global stage. The next chapter of this series will explore that aspect in more detail.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

In the end, I decided to keep the macro monitors — but publish them quarterly instead. I believe that looking back to track broader macro trends is just as valuable as zooming in on specific companies or themes.

Let me know in the comments: is this kind of content interesting or a bit boring? I’m still experimenting with the format, and your feedback genuinely matters.

Over the past few quarters, the US has pulled even further ahead of China and Europe (including the UK and Israel) in tech funding. In Q2 2025, US startups raised a massive $80B, over 5x more than their counterparts in China or Europe. That spike was largely driven by mega deals like Meta’s $14.3B investment in Scale AI and Crusoe’s $11.6B round.

India’s tech scene is gaining steam, but it's still playing catch-up—raising about $5B this quarter, which is almost three times less than China or Europe.

What’s interesting is that while the US led in funding volume, deal activity surged on both sides. Chinese startups closed around 1,900 deals, nearly matching the 2,300 in the US.

That gap between deal count and total capital raised is telling—Chinese startups are proving to be a lot more capital efficient than their American peers.

By number of deals, China leads deep tech investing with a 68% share, especially in areas like AI, compute, space, biotech, energy, and robotics. Singapore and Australia are on par with the US and Europe when it comes to deep tech investment activity, while South Korea, Japan, and India are still trailing.

In terms of total funding, India and Australia’s high deep tech share was mainly driven by a few big energy-related deals—like Greenko Group’s $1.4B secondary round, Exide Energy Solutions’ $421M raise, and Pacific Energy’s $240M PE round.

Looking closer at deep tech investment across APAC, China still dominates most categories—especially robotics and compute, where it makes up 96% and 93% of the deals. As mentioned earlier, India stood out in climate thanks to a few massive energy rounds, but it’s also starting to catch up in AI and space tech.

China still leads in number of deals, but AI and energy are starting to see more diversity, with India, Singapore, and Australia quickly gaining ground.

Since Q4 2024, AI investment in the US has completely outpaced the rest of the world. Over the past 9 months, around 90% of all AI software funding went to US companies, driven by massive rounds like OpenAI’s $40B raise (the biggest private round ever), Scale AI’s $14B, Crusoe’s $11.6B, and more. In Q2 2025, Europe edged ahead of China in total AI funding ($3B vs. $1.8B).

At the same time, the number of AI startups raising capital in China is growing faster than in both Europe and the US. A few notable examples include:

• 🇨🇳 $140M Series A from Chinese government funds: United Imaging (联影智能) is a medical AI company enabling smart equipment to support biomedical research and automating clinical workflow.

• 🇸🇬 $75M Series B from Benchmark: Manus AI, a versatile AI agent enhancing productivity and research across various domains. Media positioned it as Benchmark’s first deal in China, although the US fund invested in Manus’ Singaporean entity.

• 🇨🇳 $55M Series C+ from SingTel, Bosch and Gentree: Whale AI (帷幄), a martech startup providing brands with data and AI-driven insights and operation automation, helping them reduce costs and increase efficiency in marketing and operation.

• 🇨🇳 $42M pre-A from HongShan, Yunqi, PnP, Prosperity7: Noematrix (穹彻智能), a spin-off from another deep tech startup Flexiv, developing LLMs and infrastructure for embodied AI applications and general robotics solutions in different industries.

• 🇦🇺 $24M Series B from Bessemer: Relevance AI, an AI agent operating system that helps companies automate repetitive reasoning tasks.

Over the past couple of quarters, American compute startups have been pulling in more and more capital, driven by surging demand from AI companies—as shown in the previous chart. In contrast, compute funding in Asia and Europe has been on the decline. In China, it dropped from $4.2B in Q4 2024 to just $1.2B in Q2 2025.

That said, China still leads by a wide margin when it comes to the number of compute startups raising money—almost 4x more than the US and Europe. The number of deals in China jumped from 155 in Q4 2024 to 219 in Q2 2025, driven by strong market demand and government efforts to build a full local semiconductor supply chain. A few standout investments include:

• 🇨🇳 $224M PE round: Yangtze Memory (长江存储), the leading developer and manufacturer of 3D NAND flash memory wafers, chips and other solutions used in mobile communications, consumer electronics, computers and data centers.

• 🇨🇳 $140M Series C from state funds: Axera (爱芯元智), a developer of AI perception and edge compute chips with multi-threaded heterogeneous multi-core design which supports leading LLM/VLMs.

• 🇨🇳 $14M round from state funds: Enflame (燧原科技), an AI startup developing cloud-based deep learning chips for AI training. The company raised a $280M Series D round in 2024 and is now preparing for an IPO.

• 🇨🇳 $7M Series C from IDG and Tencent: Jaguar Micro (云豹智能), a startup specialising on programmable data processing units (DPU). Its co-founder Sunny Siu (萧启阳) created RMI that was acquired by Netlogic / Broadcom in 2009.

• 🇯🇵 $7M round from Honda Motor: Rapidus, a semiconductor company developing advanced logic chips of 2 nanometers or below and building an advanced foundry in Japan.

• 🇰🇷 $5M round from Korea Development Bank, SBI Korea: ENERZAi, an AI startup developing AI inference optimisation engines.

In Q3 2024, robotics and autonomous vehicle (AV) startups in both the US and China pulled in record-breaking funding—but that was largely thanks to two massive deals: Waymo’s $5.6B Series C in the US and a $3.2B round by Yinwang Intelligence (引望智能), a Huawei spin-off focused on AVs.

Since then, funding for Chinese robotics startups has been climbing steadily, while in the US it’s mostly flat. In Europe, robotics startups are raising more each quarter too, but the total is still 4–5 times lower than in China.

Some of the recent robotics investments in China and the of Asia include:

• 🇨🇳 $420M round from Ant Group and CATL: Zaofu Intelligent (造父智能), a robotaxi spin-off from bike-sharing platform Hellobike, focusing on developing and commercial application of L4 autonomous driving technology.

• 🇨🇳 $154M Series B from JiYuan (former GGV), CATL: Galbot (银河通用机器人), a leading company in embodied multi-modal large-model general-purpose robotics, focusing on the global customers.

• 🇨🇳 $98M Series C+ from Tencent, Alibaba, China Mobile: Unitree (宇树科技), a leading developer of general-purpose legged/humanoid robots and dexterous mechanical arms for consumer and industry applications.

• 🇨🇳 $91M Series C from Gaorong and TH Capital: Flexiv (非夕), a developer of human-inspired general-purpose adaptive robots for industrial automation.

• 🇨🇳 $91M B+ Series A from Baidu, EQT and CDH: Zelos (九识智能), a startup developing and producing L4-level unmanned robovans for urban, industrial and warehouse logistics; used for express delivery, supermarket retail, brand logistics, and community group buying.

In Q2 2025, climate investment in the US saw a rebound—even with the Trump administration’s unfriendly stance toward sustainability. Both the number of deals and total funding went up, thanks to major rounds like TerraPower ($650M), Takanock ($500M), and Silicon Ranch ($500M).

Meanwhile, climate funding in Europe keeps trending down, though the number of deals is still on par with the US.

China also had a similar number of climate startups raising money in Q2 2025. But for the first time, India pulled ahead of China in total PE/VC climate funding, boosted by big rounds for Greenko, GreenLine, and a few others.

• 🇮🇳 $1.4B secondary transaction by AM Green Group: Greenko Group, a renewable energy company, replacing fossil fuels with integrated decarbonized energy and grid assets.

• 🇮🇳 $275M round: GreenLine Mobility, a company producing environment-friendly trucks powered by liquified natural gas (LNG).

• 🇦🇺 $78M Series D from SLB: RayGen, is a startup developing integrated energy solutions combining highly efficient solar technology and long duration energy storage.

• 🇮🇳 $58M PE round from ValueQuest: Jupiter International, a leading developer of solar panels, PV models and solar energy power solutions.

• 🇨🇳 $42M pre-A round from Ant Group: ZhongHai Storage (中海储能), a company producing large scale energy storage equipment and solutions based on iron-chromium flow batteries (ICFBs).

2024 was the year China overtook the US in PE/VC funding for space tech. But in the first half of 2025, American space startups have been much more successful at raising capital compared to their Chinese counterparts.

That said, the trend flips when you look at the number of deals—more Chinese companies closed funding rounds, which follows a similar pattern we’ve seen in other sectors.

India is also quickly gaining ground on China and Europe when it comes to space tech funding.

Here are some standout space tech deals in APAC from Q2 2025:

• 🇮🇳 $100M Series B from General Catalyst: Raphe, a leading manufacturer of high-efficiency drones leveraging ultra-light carbon fiber composites technology and autonomous navigation systems powered by simultaneous localization and mapping (SLAM).

• 🇨🇳 $42M Series B+ from Bohua Capital: Qianxun (千寻位置), a satellites startup, leveraging China’s Beidou system and over 6,000 GNSS augmentation stations to provide spatio-temporal intelligent services featuring centimeter-level positioning, millimeter-level perception and nanosecond-level timing.

• 🇨🇳 $42M Series A from Plum Ventures, Bohua Capital: Ubinexus (星移联信), a satellite internet startup designing communication load systems and building an intelligent constellation networking and satellite data operation map.

• 🇨🇳 $14M Series A from Bohua Capital: SpaceIoT (鹏鹄物宇), a satellite internet full stack solution provider focusing on internet-of-things applications in maritime, mining, utilities and other industries, as well as on customer use.

• 🇸🇬 $6M round from Paspalis, NTT Finance: Transcelestial, a laser communications startup that provides point-to-point wireless communication network.

India is shaping up to be the hottest Asian market for deep tech investors from the US and Europe—especially in AI, climate tech, and increasingly, space.

EQT stands out as the most active Western backer in Asian deep tech and remains one of the few firms still consistently investing in China, despite growing caution from others.

Benchmark’s investment in Manus AI made waves across the tech world — widely seen as their “first Chinese deal.” Technically though, Benchmark backed Manus’s Singapore entity, which reflects a broader trend: many Chinese startups are now setting up HQs in Singapore to ease global expansion and navigate geopolitical headwinds.

Other examples of Singapore-headquartered startups with deep Chinese roots that raised from top-tier international VCs include Airwallex and PatSnap—both of which have successfully scaled beyond APAC.

In Q2 2025, the US remained the primary target market for Asian investors seeking deep tech opportunities outside their home region. Several funds also maintained exposure to Europe—such as Antler, Temasek, FasterCapital, QIA, and Vertex.

Among early-stage (pre-seed) investors, Antler (Singapore), FasterCapital (UAE), and Elevation Capital (India) stood out as the most active Asian backers of deep tech startups in the US and Europe, with a strong focus on AI and climate tech.

On the mid-to-late stage side, Prosperity7 (backed by Saudi Aramco) and Temasek (Singapore) emerged as the leading APAC-based global investors. Notably, both firms also maintain significant exposure to China's deep tech sector.

While VC exits in the US are slowly recovering, the situation in Asia still remain challenging. Both China and India significantly underperform in terms of the number of VC-backed M&As and IPOs.

Q2 2025 saw a strong spike in exit amount in China, but it was driven by a single transaction — $150B IPO of CATL on the Hong Kong stock exchange, over 75% of China’s total exit value in that period. If we exclude CATL, the exit amount in China would be stagnating.

At the same time, by number of exits China saw a 47% growth in PE/VC-backed tech M&As and a 32% growth in tech IPOs in the first half of 2025 compared to the same period in 2024. Although it’s a positive sign, the size of exits is a more direct indicator of the ecosystem health as a whole.

India has one of the most vibrant IPO markets in the world, there were 108 IPOs in India in Q2 2025, however, only less than 5% of the Indian tech companies going public are VC/PE-backed.

Below are some of the exit highlights from APAC:

• 🇨🇳 IPO at a $150B market cap: CATL (宁德时代), the world’s leading manufacturer of EV batteries and integrated battery solutions, major customers include Tesla, Volkswagen, Ford, Mercedes-Benz, and BMW.

• 🇨🇳 $2.5B acquisition by Tencent Music: Ximalaya (喜马拉雅), an online audio sharing platform accumulating 340 million audio contents across 100+ categories, powered by an AI voice assistant. Exit for Sierra Ventures, SIG, KPCB, Primavera, Hermitage and other funds.

• 🇨🇳 IPO at a $1.6B market cap: Unisound (云知声), an AI startup specialising on IoT solutions with focus on voice technology for smart home, healthcare, education, automotive and other use cases.

• 🇮🇳 IPO at a $1.4B market cap: Ather Energy, a manufacturer of electric scooters and provides fast electric vehicle charging. Exit for Stride Ventures, Tiger Global, InnoVen, GIC, NIIF.

• 🇨🇳 $831M acquisition: Source Photonics (索尔思), a global supplier of optical communication technology, the solutions and products for data centers, metropolitan, network communications and data connection, got acquired by an electronics equipment manufacturer Dongshan Precision (东山精密).

• 🇨🇳 IPO at a $765M market cap: Breton (博雷顿), developer of autonomous electric trucks for construction, mining and other industries.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn.

Let’s connect and explore how we can improve together.

Building global businesses is hard. Consistently creating multiple global champions across different industries? Almost impossible.

Unless you're China.

In just two decades, Chinese manufacturers in sectors like solar panels, wind turbines, and electric vehicles (EVs) have gone from near zero to 60–80% of global market share.

It looks like China has discovered a “secret sauce” for scaling industrial giants. But is it truly sustainable? And can it be replicated in other industries?

China’s “Magic” Formula

All these three industries in China — solar panels, wind turbines and EVs — have gone through the same evolution:

Stage 1. Inception. Key technologies were transferred to China by global market leaders in exchange for access to its massive and fast-growing domestic market. Companies like Applied Materials (PV manufacturing equipment), Vestas and Gamesa (wind turbine designs), and global automakers (Volkswagen, GM, and Toyota) agreed to share or localise technologies through JVs or licensing agreements with Chinese partners (SAIC, FAW, Dongfeng etc). These firms were lured by the enormous revenue potential of what was becoming the world’s largest market for clean energy and autos.

Even Tesla launched Gagifactory in Shanghai. Although it was a wholly-owned entity, some diffusion of talent probably may have taken place.

Stage 2. Strategic guidance. After noticing a promising tech trend, China’s central government designates this direction as the national strategic priority and includes it into the next Five-Year plan. This commands a clear signal to local governments to start providing generous incentives to any companies that are involved in strategic sectors. The toolsets are very extensive: subsidies, grants, tax breaks, cheap land, low-interest rates, procurement mandates or local component requirements.

In some cases, government support began almost immediately after the emergence of the first domestic players — for example, local content requirements for wind turbines were introduced as early as 2006, and consumer subsidies for electric vehicles rolled out in the early 2010s. In other cases, such as solar panels, the government stepped in only after domestic firms had already captured a dominant global position: by the time the Golden Sun Program was launched in 2009, Chinese manufacturers had already secured over 50% of global market share.

Important to note that by taking these “benefits”, companies also bear heavy responsibility to perform according to the strict KPIs (revenue growth, profitability, taxes etc).

Stage 3. Oversaturation. Such favourable conditions give birth to hundreds of copycats that flood the domestic market with cheap barely differentiated products. For example, in mid 2010s there were around 500 EV brands and thousands of solar panel producers in China. All this contributes to creating a broader industrial ecosystem with the complete and interconnected supply chain within one country or sometimes even within one province.

Stage 4. Price wars. Competition between local companies on the domestic market intensifies, prices fall, margins are being squeezed, many companies die. Some of them fail to meet government KPIs and get cut off from the subsidies. For example, 400 Chinese EV companies ceased operations between 2018 and 2025.

But for China, this is not a flaw, but a feature — both success and failure contribute to building the supply chain. As venture capitalist David Li aptly puts it: “China builds arenas for the Hunger Games.”

Stage 5. Natural selection. Such a harsh market environment forces Chinese companies to innovate faster and increase operational efficiency. Eventually, those with the best unit economics and largest cash war chest survive and start expanding into the global markets where margins are higher. The domestic market starts consolidating.

Stage 6. Global dominance. The survivors emerge as national champions, expand into global markets, and outcompete international rivals through aggressive pricing, competitive product quality, and deep cash reserves. Ironically, the casualties of China’s expansion are often the very companies that once transferred the original technology—like Vestas, Gamesa, and Volkswagen.

Stage 7. Price wars continue. Even after acquiring a major market share, the Chinese players continue fiercely competing with each other, which eventually leads to overproduction, both in China and globally. Margins keep falling, national champions become loss-making and force their local government sponsors come for rescue (too big too fail), as it happened with NIO being saved by the Anhui government in 2024.

Stage 8. Equilibrium. At some point, price wars end through setting price floors, either by the government intervention (solar panels) or agreements between the largest players (wind turbine). For EVs, such a solution is still to come.

State-Driven Model — Key Success Factors

Why did these three industries in China manage to become so competitive both on the local and global markets in such a relatively short period of time, while other countries fail to catch up?

Below are the key factors sorted by their contribution to success:

1. Government support

The central government, along with provincial and municipal authorities, implement targeted measures to support solar panel, wind turbine, and EV manufacturers. This support mainly focuses on 1) enhancing the cashflow reserves (direct subsidies, grant, low-interest loans), 2) optimising cost (R&D subsidies, land grants), and 3) boosting revenue growth (feed-in tariffs, consumption subsidies, local component procurement requirements etc).

In the solar sector, the "Golden Sun" program (2009–2013) provided billions of RMB in subsidies for PV deployment. Similarly, wind power benefited from the Renewable Energy Law of 2005 and subsequent feed-in tariffs that guaranteed fixed returns.

For electric vehicles, the government rolled out direct consumer subsidies of up to $10,000 per vehicle in the early 2010s, along with extensive public investment in charging infrastructure. Between 2009 and 2022, total government subsidies for EVs exceeded $30B. BYD alone received almost $2B subsidies from various local governments between 2020 and 2022.

2. Talent

China's extensive higher education infrastructure and focus on STEM education have yielded a vast pool of engineering and technical talent. As of 2023, China graduates more than 1.4 million engineers annually, the highest in the world. Universities such as Tsinghua, Zhejiang, and Shanghai Jiao Tong play major roles in clean energy R&D, often in collaboration with industry partners.

For example, Huawei’s ICT Academy program has partnered with over 3,000 universities globally, training more than 1.3 million students in fields like AI, cloud, and 5G. Other programs are pushed by the government, such as the joint-talent training initiative of the Ministry of Education which has linked more than 2,400 employers with more than 2,000 universities, benefiting more than 6.2 million students in strategic tech areas (AI, semiconductors, robotics etc).

In terms of patents, China leads global filings in solar PV, wind turbine design, and EV battery technologies. According to WIPO, after 2020 over 30% of global patents in solar and other energy-related technology are filed by Chinese entities. Leading firms like CATL (for EV batteries) and LONGi (for high-efficiency solar wafers) invest over 5–7% of annual revenue into R&D.

3. Innovation

China’s green tech rise has been accelerated by both internal innovation and the strategic transfer and adaptation of foreign technologies—often facilitated by state policy. In solar, the early 2000s saw Chinese firms import manufacturing equipment from Germany and the US, while hiring experienced engineers from Taiwan and Japan. The 2008 financial crisis triggered a wave of bankruptcies among Western PV firms, enabling Chinese companies to acquire equipment, patents, and technical talent at a discount—rapidly localising and scaling solar manufacturing.

Wind turbine development followed a similarly state-guided trajectory. During the 2000s, China imposed mandatory technology transfer requirements on foreign wind firms in exchange for market access. Firms such as Goldwind and Mingyang licensed turbine designs from Vensys (Germany) and Aerodyn (Denmark), then gradually developed proprietary solutions. China’s National Development and Reform Commission (NDRC) reinforced this by tying project approvals and feed-in tariff eligibility to local content thresholds — a policy that significantly accelerated domestic technology absorption.

The electric vehicle (EV) sector saw a more refined and targeted approach to tech transfer, driven by lessons from the traditional automotive industry. In the 1990s and 2000s, foreign automakers like Volkswagen, GM, and Toyota operated via joint ventures with SOEs such as SAIC and FAW, but retained control over core technologies like internal combustion engines and transmissions—components that were not subject to local production mandates. Consequently, China failed to domesticate critical elements of traditional automotive manufacturing.

In contrast, China’s EV industry was shaped by stricter localization mandates, including requirements that vehicles use domestically produced batteries and motors to qualify for government subsidies under the New Energy Vehicle (NEV) policy regime. This policy gave domestic players like BYD and CATL a significant head start, enabling them to gain a competitive edge through vertical integration of battery production and powertrain systems. Foreign automakers were effectively constrained unless they partnered with Chinese firms or established local R&D and manufacturing operations. While these requirements were lifted in 2018, allowing Tesla to enter the market in 2019 with its wholly foreign-owned Gigafactory in Shanghai, the domestic EV ecosystem had already matured. By then, Chinese brands had secured dominant positions across the EV value chain—from cells and packs to drive units and final assembly—making them formidable global competitors.

4. Supply chain integration

A unique feature of China is that it has built the world’s most complete and cost-effective supply chains for solar, wind, and EV manufacturing. In solar, China controls over 80% of the global supply of polysilicon and wafers with major hubs in Jiangsu and Sichuan provinces that produce nearly every component in-house, reducing reliance on imports and lowering costs through scale.

For wind, China produces most of the nacelles, gearboxes, blades, and inverters domestically. Integration between state-owned utilities, component suppliers, and OEMs like Goldwind and Envision streamlines production and deployment. In EVs, China leads the world in lithium-ion battery production, accounting for over 75% of global cell manufacturing capacity in 2024. CATL and BYD anchor domestic EV supply chains that include battery metals refining, anode/cathode materials, and electric drivetrain components.

5. Domestic market

A massive home market has provided the runway for scaling production and achieving learning curve efficiencies. In solar, domestic deployment exceeded 887 GW by the end of 2024, making China the largest PV market globally. In wind, China installed over 80 GW of new capacity in 2024 alone—more than the rest of the world combined. EV sales in China reached 11 million units in 2024, representing over 60% of global EV sales.

Usually having a large domestic market discourages players to come out and fight with the global competition in a more difficult environment. However, with China it’s different, because surviving on the local market is already hard enough and global expansion is seen as one of the ways to scale and even survive.

6. Low prices

The key selling point of most Chinese brands is that they are priced lower than their American and European rivals, while maintaining competitive performance and functionality. For example, BYD and XPeng are usually 30-50% cheaper than comparable models of VW and Tesla, Chinese onshore wind turbines 20-30% cheaper than Siemens or Vestas.

Having said that, low prices do not always mean lower cost. Chinese wind turbine producers Goldwind and Envision have 4-6% gross margins, while for Vestas and Nordex 12-20%. In solar, China’s Jinko Solar gross margin in 2024 was 10.9%, while for the largest US solar company First Solar it was over 40%. BYD and Tesla gross margins for comparable car models (Tesla Model 3 and BYD Han EV) were similar, around 22-25%.

Among all these factors, government support is pretty crucial. It’s not only about subsidies and tax breaks — other countries also introduced supporting measures for green energy and other domestic sectors (e.g. Europe). The uniqueness of China is that its government support enabled other factors like talent nurturing, tech transfer and supply chain integration.

It’s probably one of the reasons why India couldn’t achieve the same result in creating competitive high-tech industries, while having a comparably large domestic market and strong talent pool.

The same state-driven model applies to many other emerging industries in China that are considered to be “strategically important”, such as AI, robotics, space, semiconductors, quantum computing.

Exception — Consumer Internet. Bottom-Up Approach

One of the brightest exceptions are companies like Alibaba, Tencent, Bytedance, Pinduoduo and other consumer internet giants. These companies didn’t take any significant subsidies from China’s government for their core businesses (not talking about their cloud divisions), but were rather funded by VC money.

There are two reasons for this:

1. Not in the Five-Year plan. Although stimulating consumption is also a priority for the Chinese state, its efforts mostly focused on the domestic market. Creating global gaming or e-commerce giant doesn’t provide the same strategic edge in the geopolitical race against the USA, as AI or space technology. At least, it was the case before TikTok became so huge in the US.

2. Easy to raise VC money. Consumer internet in 2000-2010s was a very hot topic among private investors both globally and in China, so it was relatively easy to raise mega VC rounds from Softbank, DST, Tiger others. Meanwhile, strategic areas are more long-term and did not necessarily fit into VC’s mandate back then.

Technological innovation wasn’t the decisive factor behind the global success of companies like Alibaba, Shein, or Pinduoduo. They managed to successfully compete with Amazon or other global e-commerce giants not through cutting-edge tech, but rather through scrappy execution, deep integration with Chinese supply chains, massive domestic scale, and ultra-competitive pricing.

Tencent’s dominance in global gaming similarly wasn’t built on original IP, but on the back of China’s vast user base, abundant capital, and a brilliant M&A strategy —acquiring or investing in leading studios like Riot, Supercell, and Epic Games.

The clearest exception is ByteDance, which did distinguish itself through its unique recommendation algorithms powering TikTok’s feed. But even in ByteDance’s case, execution speed, product iteration, and aggressive marketing were just as critical to its global breakout.

Generally, the evolution of the consumer internet companies is also pretty much the same as for the state-driven industries described above — harsh domestic competition (e.g. apart from Alibaba, China had over 100 other e-commerce platforms in mid 2000s), price wars, squeezed margins, global expansion, beating the competition there, reaching 20-50% global market share.

However, many other success stories were state-led.

Failure cases — Cloud, SaaS, Semiconductors

Of course, Five-Year plans don’t always work as planned.

Areas like semiconductors, cloud infrastructure and B2B SaaS applications have been supported by both local governments and VC/PE funds, but still fail to reach significant success outside of China.

There are multiple articles by , and other brilliant authors explaining why Chinese cloud giants lag behind Amazon and Google (e.g. here), why China doesn’t have its comparable versions of Salesforce or ServiceNow (e.g. here, here and here) or why China’s SMIC still struggles with its 7 nm nodes despite raising over $5B in equity and subsidies.

We can summarise these answers based on the framework of factors introduced above:

1. Semiconductors. Although the talent is there and the domestic market is huge, Chinese foundries still focus on producing low-end chip, because they still depend on ASML’s EUV lithography and have limited access to EDA tools from Synopsys or Cadence and other advanced equipment from LAM Research and Advanced Materials due to the US export controls. It significantly slowed down the development of high-end supply chain in China.

2. B2B SaaS. The biggest challenge is that the actual B2B SaaS market in China is pretty small. It’s only the 7th largest market in the world ($9B), similar size as Australia and smaller than France, Canada and Germany. SME customers usually don’t have a lot of spare cash to invest because of generally low margins and automation is still not the most pressing issue due to low labour costs. Enterprise customers follow the “build-over-buy” principle, considering that they have access to relatively cheap and good quality developers and require high customisation for every implementation.

3. Cloud Infrastructure. This market is also not very big in China (AWS generates more revenue than all the Chinese cloud companies combined), which is the function of the small SaaS market. As explained in his article, in contrast to the US, Chinese cloud providers generate their revenue mainly from low margin IaaS (Infrastructure-as-a-Service), because in China customers treat the cloud more as a commodity rather than a platform to transform business operations and enhance productivity.

It’s important to note that these “failure cases” can still be turned around with the new cycles of technology.

The same happened with automotive industry — when I wrote my master thesis in 2018 about tech transfer examples in China, automotive was still considered as a failure case. However, with the successful adoption of EV batteries, Chinese automakers managed to ride the wave of EVs and increase its share on the global automotive market from 10% in 2014 to 23% in 2024.

The same may potentially happen in semiconductors if China develops its own advanced lithography or successfully masters the next generation compute architecture, e.g. optical computing or quantum computing. Similarly, AI adoption may disrupt the existing limitations of the cloud and B2B software markets in China.

Can Emerging Industries Replicate the Success?

Obviously, this section is a bit of a look into the crystal ball, but still based on the framework above, I’ll sort the industries by their success probability and try to make a few educated guesses:

1. Robotics. China is already a leader in industrial robotics and it has all the factors to win over the race in autonomous vehicles, humanoid robots and other emerging robotics areas. It has an almost complete supply chain (sensors, actuators, mechanical structures, power electronics, software) and huge domestic market as the world’s largest production base. Chinese robotics companies will most likely replicate the global success of EVs, although there will be limitations for entering sensitive areas in developed countries (defence, finance, healthcare).

2. Logical & memory chips. Huawei and some emerging players (Enflame, MetaX etc) are making considerable progress in developing the next generation GPUs and CPUs that can already compete with NVIDIA on the Chinese market. The US export controls and still underdeveloped semiconductor production ecosystem in China are the main hurdles for further R&D advancements. If these challenges are solved either through more successful tech transfer or through adopting the next generation platform (photonics, quantum), leveraging the existing expertise of selling hardware globally won’t be too hard.

3. LLMs and AI dev tools. China has definitely made a stir with DeepSeek and other breakthrough AI models which by some parameters are even more superior than OpenAI or Anthropic. A lot of success is in open source. However, so far we’ve seen success only in the consumer segment (Minimax, Question.ai etc), while scaling in the B2B segment still remains a challenge. To win a significant share on the global markets, the Chinese developers of AI infrastructure tools need to master the B2B SaaS sales process, which is the biggest bottleneck.

4. Satellites and Rockets. Chinese state-owned satellite manufacturers and space tech startups already make the first success in expanding into developing countries (Southeast Asia, Africa). However, as Blaine Curcio mentioned in his article, the satellite market is nowadays much smaller than the EV market and it’s much more influenced by geopolitics, which makes scalability a bit challenging.

Final Remarks — Role of Venture Capital

As the table below illustrates, all of the Chinese industries mentioned in this article had VC/PE-backed leaders (at least one). Despite massive subsidies, industry champions are still eager to take VC money. It is especially noticeable for the new industries, like AI, robotics, space. But even at the early days of solar panels in China, the first investment rounds of Jinko Solar were led by US funds (Flagship Pioneering, CIVC Partners).

The three groups of companies outlined above operate under distinct motivations and face different challenges:

• VC/PE-backed firms are driven primarily by financial returns and investor exits. They scale aggressively and often lead in innovation, but rely heavily on external funding to support high burn rates, making capital access a constant pressure point.

• Subsidy-backed private companies focus on building cash-generating, scalable businesses, while aligning closely with government KPIs. Their success depends not just on market performance, but also on maintaining policy alignment and securing ongoing support from state programs.

• State-owned enterprises (SOEs) are less focused on short-term profitability and more on strategic ecosystem building—supporting domestic supply chains, standard-setting, and industrial resilience. Their key challenge is not financial sustainability, but rather overcoming bureaucratic inertia and inefficiencies that can hinder speed and adaptability.

Despite these structural differences, category leadership can emerge from any model. For example, Jinko Solar, the world’s largest solar panel manufacturer, is VC/PE-backed; Goldwind, the global leader in wind turbines, is state-owned; and BYD, a dominant EV player, is a private firm whose growth was supercharged by government subsidies.

For the perspective of a VC investor, betting on a potential leader in an emerging strategic sector in China may be a viable strategy. But in this case, the key challenge is to pick the VC-backed startup that can survive in the uphill battle with the much better funded state-owned or private behemoths.

In other words, how to differentiate between XPeng (or Li Auto, NIO) and 400+ other EV brands that died in the last seven years?

And of course, how to mitigate all the legal risks related to such investment?

But this is perhaps a topic for another article…

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

In the last decade, China has become the second dominant player in space, rivaling the US with its ambitious plans in exploring the Moon and beyond, as well as building the next generation space-enabled economy. China holds a solid second place in terms of orbital launches, it has its own space stations and manned spacecrafts, and is rigorously working on bringing the first taikonauts to the Moon by 2030.

The space industry in China – as any other major superpower – has historically been shaped through its massive government programs led by state-owned enterprises like CASC, CASIC and China Satcom, as well as government agencies like CNSA and SASTIND.

Although China does not officially disclose its space budget, research firms like Euroconsult estimate that Chinese government space spending reached approximately 20 billion dollars in 2024, making it the second-largest in the world after the United States (around 80 billion dollars). It is also nearly three times larger than the 7.7 billion euro budget of the European Space Agency (ESA).

China’s space budget has grown significantly—from 12 billion dollars in 2022—and is expected to continue rising, particularly with the expansion of its lunar program, which some forecasts suggest may require over 100 billion dollars in total funding.

In 2014, China opened its aerospace industry to private investments, which has eventually created a vibrant ecosystem of Chinese private space tech startups.

Fast forward to 2024, China for the first time surpassed the US in terms of VC/PE funding for space tech startups ($2.7B vs $2.6B). However, the Chinese private space industry still remains a bit of a mystery for almost everyone outside of China.

In this article we’ll try to shed some light on it and explore if there are any ways for global investors to leverage this booming sector in China.

State-Driven Venture Capitalism in Space

For this research, we analyzed 130 space tech startups from the Chinese VC database ITjuzi, including 85 VC/PE backed startups.

Private space sector in China has evolved through two major waves of development:

1. Pre-COVID wave (2015-2020): early success stories like pioneering rocket companies i-Space, Landspace and ExSpace, as well as satellite producer ChangGuang, driven by regional funds with an active participation of private VC firms such as Matrix China, Sequoia China, and Shunwei Capital.

2. Post-COVID wave (2021 - present): supercharged by investments from numerous regional funds and state-owned banks that backed such new leaders as satcom unicorn SpaceSail and rocket makers: SpacePioneer and Galactic Energy.

2021 marked a fundamental shift in the Chinese VC industry from consumer-oriented startups towards the next generation of deep tech projects, including space tech. This trend has been massively driven by the Chinese government and has ultimately reshaped the whole VC landscape in China. During the second wave, the share of direct government funding into space tech startups increased from 20-30% to 60-70%.

The significant involvement of the Chinese government has contributed to the perception that the country’s space industry operates as a closed ecosystem—largely inaccessible to foreign investors and suppliers.

But is that perception entirely accurate?

The Great Space Wall of China

To answer this question, we need to understand the structure of the Chinese private space industry in more granular terms.

We can divide all Chinese space tech startups into three major categories:

1. Manufacturing & Launches. Most manufacturers of rockets and satellites in China – same as SpaceX in the US – design, assemble and launch their vehicles in-house. These companies are served by a whole ecosystem of vendors that supply electronics, propulsion systems and other auxiliary components, while most key components like engines are still manufactured by CASC.

2. Operations & Services. These companies provide control and tracking systems for satellites, as well as communication infrastructure between space and ground systems.

3. Applications. This segment includes the operators of navigation, remote sensing and communication satellite constellations, as well as emerging deep tech startups, e.g. providers of nano satellites, edge compute infrastructure in space and others.

There are no explicit restrictions for Chinese private companies to invest in any of these segments, however, for some military-related projects they need to obtain certain approvals.

Currently, around 40% of total funding for Chinese space startups originated directly from government funds, while according to some expert estimations another 40% may be contributed indirectly via LP relationships with private VC/PE funds. The rest is coming from private investors (insurance companies, family offices and other LPs).

The largest space tech segments in China are rocket and satellite manufacturing, as well as satellite communication services. All these segments have a relatively large share of government funding (40-50%). The Operations & Services segment is also dominated by state-owned funds (50-70%), which may be explained by potentially low margins of these businesses and hence less interest from commercial investors. Emerging technologies also have high government ownership due to their connection to state-owned labs.

Meanwhile, segments such as remote sensing, navigation-related services, and the manufacturing of satellite and rocket components are increasingly led by private investors.

Success Stories of Chinese Space Startups

In the last five years, there were four notable space tech exits in China:

1. Piesat – satellite data provider for 3D visualization. Went public in 2019 on Shanghai Stock Exchange at a $1B market cap. Exit for QF Capital.

2. Geovis – providing the 3D mapping and data platform of the Earth for defense, mining, climate and other applications. The product includes an open API platform, data sets and 3D modeling tools for developers. The company went public in 2020 on Shanghai Stock Exchange at a $2.8B market cap. Exit for Sugon.

3. Everlight Space – satellite energy systems provider. Secondary transaction in 2024: exit for Oddor Capital.

4. Teemsun – high-precision navigation satellite manufacturer. Went public in 2024 on Shenzhen Stock Exchange at a $280M market cap. Exit for Lenovo Star.

There should also be significant markups for early-stage private investors that backed the following companies:

Landspace – manufacturer of liquid oxygen and methane fuelled rockets:

• 2016: FounDream invested at a seed stage

• 2020: Raised a $240M Series C from Matrix, Hongshan and Lightspeed

• 2024: Raised another $126M from a state-owned fund

i-Space – manufacturer of cost-effective commercial rockets:

• 2017: Shunwei, Fosun and CITIC invested at a ~$4M seed round

• 2018: Matrix led a ~$60M Series A

• 2024: $100M Series C by state owned funds

MinoSpace – turnkey manufacturer of communication satellites:

• 2017: FounDream, Essential Capital and CAS Star wrote a $1.5M seed check

• 2021: Lightspeed led ~$40M Series B

• 2024: Oriza Holdings led a $140M Series C

The Universe of Chinese Space Investors

The model of specialist VC funds is not common in China. That is why, in contrast to the US or other countries, China doesn’t have space-focused funds, like Seraphim, Space Capital or Beyond Earth Ventures. Instead, it has multi-specialist deep tech funds (both private and state-backed), as well as generalist funds that developed their space practice in-house.

There are really strong state-backed deep tech funds in China that actively invest in space, e.g. CAS Star under the Chinese Academy of Science, Shenzhen Capital Group or Addor Capital under Jiangsu High-Tech Investment Group.

Among TOP tier funds, Matrix China is by far the most advanced in terms of space tech investments, while Hongshan is much more generalist. At the same time, there are several Tier-2 deep tech funds with strong space tech expertise, e.g. Essential Capital or Hongfu Investments.

Another peculiarity of venture investments in China, is that the borderline between VC and PE mandates is extremely vague. Both private and public PE funds (Jiangsu Jinyu, Founder H Fund etc) and even asset managers (CICC, BSCOMC, etc) may invest in Series A or even seed startups.

Can Foreigners Invest in Chinese Space?

Instances of foreign investment in Chinese space startups are rare. One notable case is the investment by Abu Dhabi's state fund, ADIA, in the Chinese rocket manufacturer Ming Kong (明空航天) in December 2023. However, this investment is widely viewed as a strategic political move rather than a purely financial decision.​

Additionally, while top-tier global funds like Sequoia, Matrix, and Lightspeed have active Chinese teams investing in the space tech sector, these teams operate independently and are fully localized, separate from their parent funds.

Why don’t foreign funds invest in Chinese space startups if the sector is booming?

China's regulatory framework imposes significant restrictions on foreign investments in certain segments of the aerospace sector, particularly those related to national security:

• Negative List. The 2024 version of the Negative List for Foreign Investment Market Access prohibits foreigners from investing in aerial imaging, remote sensing and navigation mapping projects.

• Dual Use. Some satellites and rockets may classify as weapons, which makes direct investment automatically impossible for foreign entities. Also, the Negative List broadly specifies that all foreign investment transactions related to national security must undergo approvals with Chinese regulatory agencies.

• US Restrictions. Apart from Chinese restrictions, US export controls also prevent American and other global investors from backing China’s strategic areas which can be used for military applications, including aerospace.

Would Indirect Investments Help Bypass These Restrictions?

If the area where the startup operates is listed in the Negative List, any investments through SPVs (special purpose vehicles) would also be restricted. For dual-use companies an indirect foreign ownership below 20% is possible, but subject to approval.

Although investing as a Limited Partner (LP) into one of the Chinese funds may help solve this problem, putting money in an offshore GP structure (outside of China) may cause additional concerns with regulators. Moreover, there is a special license called QFLP that generally allows foreign investors to back Mainland Chinese funds. However, QFLP still needs to satisfy the requirements of the Negative List.

At the same time, a few years ago all manufacturing-related activities were excluded from the Negative List and added to the list of areas encouraged for foreign investments, including the design and manufacturing of rockets, civil satellites and their components.

Ultimately, almost any foreigners investing in a Chinese space startup need to obtain approvals from Chinese regulators and make sure their investment targets are not recognised as military companies.

Due to strict oversight, geopolitical tensions, and regulatory uncertainty, many overseas investors have understandably stayed away from China’s space tech sector. At the same time, Chinese startups themselves show little interest in foreign capital, because it’s often not worth the trouble, given all the restrictions and strong demand from domestic investors.

That said, this trend may change in the future as more startups begin to prioritize global expansion. As the ecosystem matures and becomes more dynamic and open—especially in areas like manufacturing—it may become more accessible. This may prompt global investors to start to learn about the Chinese ecosystem and find ways to navigate the challenges in order to gain exposure to China’s growing aerospace industry.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Highlights:

• Exit value comparison. In 2024, Asian startups delivered $300B in exits, with China contributing $207B. The US led with $438B, while Europe trailed at $81B.

• Exit count comparison. The US topped exit count with 1,477 deals. Europe surpassed APAC (813 vs. 735), while China had just 449.

• IPO vs M&A. China and India are more IPO-driven, and since IPOs are 5–6x larger than M&As, Asian exits yield higher returns per deal.

• Average ROI: China 2.7x, India 3.1x — ahead of the US (2.0x) and Europe (1.2x). But exit rates are higher in the West: 14% (US), 13% (Europe), vs. 8% (China) and 6% (India).

• China’s Rise and Fall. From 2017 to 2021, Chinese startups outpaced the US in exit liquidity. But since then, China’s exit value has fallen over 80%, mainly due to an IPO slowdown — a trend that continued in 2024 even as other markets rebounded.

• China’s Back on Nasdaq. The drop in Chinese exits no longer seems tied to regulation or geopolitics. In 2024, 24% of Chinese IPOs listed on Nasdaq, up from 5% in 2023 — making up 26% of all Nasdaq listings.

• Mood is Bearish. Instead, the slowdown reflects macroeconomic headwinds, weak investor sentiment driven by poor stock market performance and general China’s policy shift from supporting IPOs in non-strategic sectors.

• Deep Tech to the Rescue. Chinese deep tech exits will increase and follow two tracks: domestic IPOs from profitable “small giants” in strategic sectors, and overseas listings by globally-focused tech leaders like Pony.ai and WeRide.

Exits are one of the key parameters that define the VC ecosystem. They are perhaps more important than the size of VC funding itself, because without a mature exit market, even vibrant VC ecosystems will sooner or later stagnate.

That’s what happened in markets like Europe or Southeast Asia, where VC funding was peaking in 2020 and 2021, but then without clear path to liquidity a lot of VC managers struggled to demonstrate high DPI (distributions-to-paid-in capital) numbers to their LPs.

Low DPI — Few LPs re-investing into new funds — Fewer / smaller funds — Less capital for the startup ecosystem — Repeat…

This is why a lot of global fund managers are re-focusing on the US, as this is the only large VC/PE market with a smooth access to liquidity through M&A and IPO.

Or at least this is the common perception.

Perception vs Data

Perception or not — you never know until you look at the data.

The map below compares countries by the number and value of VC/PE exits in 2024. Following the Pitchbook methodology, in VC/PE exits we include IPOs and M&A deals involving companies that raised at least one VC or PE round prior to exit.

The map also highlights exits in deep tech sectors such as AI, advanced computing, robotics, climate, and space.

The data proves — the US is definitely the biggest source of VC/PE exits, which rightfully makes it now the best place in the world to do venture investments. In 2024, the American PE/VC-backed companies had $438B of total exits across 1,477 deals, which is 21% more than the year before.

What about other markets?

It depends on the methodology. In quantitative terms, Europe holds a solid second place with 813 exits, which is 11% more than in the the whole APAC, including China. Over 60% of Asian VC/PE backed exits occured in China (449 exits) — nearly 2x fewer than Europe and 3.3x fewer than the US. The rest were spread across India (110), Australia (45), Singapore (33), Japan (24), South Korea (18) and others.

However, when it comes to exit value, in 2024 China was 2.5x larger than Europe, with $207B generated in exits, though still less than half the size of the US. India’s $47B in exits was just 72% below Europe’s total, despite having 7.4x fewer exits.

Exits-to-investments ratio is another useful metric for estimating the liquidity of a VC ecosystem, both by deal count and total value.

The US and Europe have higher exit-to-investment ratios by deal count, meaning startups there are more likely to reach an exit. In contrast, China and India outperform on value-based ratios, suggesting that while exits are less frequent, they tend to generate significantly higher liquidity per deal.

The US and Europe have higher exit-to-investment ratios by deal count than Asia markets, meaning that every investment in American or European startups is more likely to lead to an exit. However, China and India boast higher ratios by value, which implies that Asian startups generate on average larger liquidity amounts.

Australia appears highly liquid across both metrics. However, its 2024 exit value data is not very representative — of the $18B in exits, $16B came from a single deal: Blackstone’s acquisition of AirTrunk.

2024 marked a year of recovery, with exit activity gaining momentum in Europe (79%), the US (21%) and India (150%). In contrast, China continued its decline, with exits falling further both in volume and value.

Notably, between 2017 and 2021 Chinese startups were generating more liquidity than their US counterparts. But since 2021, China’s exit value has dropped by over 80%.

Key Drivers

For simplicity’s sake, we divide all VC/PE-backed exits into two categories: 1) IPOs, including stock exchange and over-the-counter listings, 2) M&As, including strategic trade sales and buy-outs.

The key difference between Western and Asian VC ecosystems lies in the exit pathways: while the US and Europe are primarily M&A driven, Asia — especially China — is heavily reliant on IPOs.

In terms of number of exits, IPOs have consistently made up less than 20% of all exits in the US and Europe, whereas in Asia IPOs typically account for 30-50%. In China, that share is even higher, often reaching 50-70%.

IPO exits are typically 5–6x larger than M&A exits, as listing on a public exchange requires startups to reach a significant scale — unlike strategic acquisitions, where size is often less critical.

This helps explain why China is twice as large as Europe by exit value, yet has half as many exits. The same pattern holds for India, where a smaller number of outsized IPOs drive the bulk of liquidity.

Back in 2021, the global VC boom was largely fueled by mega IPOs—so much so that startups in the US and Europe generated more liquidity from IPOs than from M&A that year, a rare shift in the typical exit landscape.

Busting Some Myths About China

This is my favourite part!

Myth #1: Liquidity crunch in China — partially true

China’s exits-to-investments ratio (2.7x) is higher than in the US ratio (2.0x). It means that the Chinese VC/PE ecosystem is still more balanced in terms of liquidity than the US ecosystem.

In 2024, China saw substantial amount of massive IPOs, including a lot of listings of deep tech startups, such as Horizon Robotics ($5.8B), Pony.ai ($4.1B), WeRide ($3.6B) and Innoscience ($3B). Some of them were even NASDAQ listings, despite the geopolitical tensions. Also there were some sizeable M&As, e.g. Shanghai RAAS was acquired by Haier and ProfoundBio was acquired by the Danish Genmab.

However, as mentioned above, China’s exit activity has been declining since 2022, both by number of exits and by exit value. It is mostly driven by the declining number of stock exchange listings: from 403 listings in 2021 to 141 listings in 2024.

Myth #2: Chinese companies can’t access overseas IPOs — false

There were almost no overseas IPOs by Chinese companies between 2021 and 2023 because of geopolitical tensions and regulatory crackdown by the PRC government. It introduced strict rules obliging Chinese companies to obtain approvals for overseas IPOs. So, a huge backlog of pending applications has accumulated since 2022.

However, in mid 2023 the China Securities Regulatory Commission (CSRC) started signalling that it would continue to unblock overseas listing channels for Chinese companies. Before that China and the US also reached an agreement to share audits of US-listed Chinese businesses, which previously caused a lot of delistings.

In 2024, 52 Chinese companies went public on overseas stock exchanges — the vast majority of them on NASDAQ. The share of overseas listings among all Chinese IPOs jumped from just 3% in 2022 to 24% in 2024. Chinese companies also made up 26% of all NASDAQ listings that year.

But despite this rise in offshore IPO activity, the overall number of IPO exits from Chinese companies continues to decline. This likely reflects deeper structural issues in the Chinese economy — from deflationary pressures and slowing growth to rising unemployment. Add to that the consistently cautious investor sentiment around China over the past few years, and the IPO slowdown starts to make more sense.

Moreover, the Chinese government has increasingly focused its support on so-called “small giants” (小巨人) — companies operating in strategic industries such as advanced manufacturing, semiconductors, and industrial tech. In contrast, many consumer-facing companies have faced delays or rejections when seeking to go public.

As a result, the share of “small giant” IPOs on the Beijing Stock Exchange has surged from 38% in 2020 to 70% in 2024, even as the overall number of listings declined — highlighting the state’s clear preference for strategic sectors.

Myth #3: China’s M&A market is stagnating — true

Can Chinese M&A market save the day? — Probably, no.

Over the last ten years, it has been stagnating. Every year, there are some sizeable strategic acquisitions announced, such as Shanghai RAAS and ProfoundBio in 2024. But M&A has never been a driving force for exits in China.

There are a few reasons for this:

• Regulatory Hurdles: China has long restricted certain M&A types — including cross-industry deals, takeovers of loss-making firms, and PE acquisitions of listed companies. New antitrust and national security rules have added complexity, often delaying or blocking deals — like the 2021 halt of Tencent-backed Huya and Douyu’s $5.3B merger.

• Limited Pool of Strategic Buyers: State-owned enterprises typically focus on acquiring assets in strategic industries or distressed firms, not VC-funded tech startups. Moreover, foreign tech companies have been cautious with acquiring Chinese startups due to China’s foreign investment restrictions.

• Cross-Border Scrutiny: Since 2021, foreign acquisitions in key sectors require national security clearance in China. At the same time, Western governments are tightening restrictions on Chinese tech through capital controls, making cross-border exits riskier.

• Cultural Aspect: IPOs are seen as the ultimate success, while selling at anything less than a sky-high price – a startup can be viewed as a failure. Many Chinese founders are emotionally attached to their companies and resist M&A — viewing it as "giving up their child" rather than a strategic win. For example, when retail mogul Huang Guangyu once offered to buy rival Suning, Suning’s founder rebuffed him by saying he’d rather give the company away for free if he failed, than sell – selling is akin to admitting defeat.

What is Next for VC Exits in China?

I posed this question to several gurus in China’s VC space, including David Li from Open Innovation Lab. The consensus is clear: in the years ahead, deep tech–focused “small giants” will play an increasingly central role in China’s VC ecosystem.

Backed by strong government support, these companies will often become profitable early and choose to stay private, rather than pursue IPOs. Take Unitree Robotics, for example — founded in 2016, it has been consistently profitable since 2020. This shift could mean fewer IPOs, but not necessarily lower VC returns, as investors may receive dividends instead of exits. A VC model — with Chinese characteristics.

At the same time, we may see a separate track for globally-oriented Chinese deep tech startups, like Pony.ai or WeRide, which have raised capital from foreign investors and gained international traction. These companies are more likely to go public abroad — likely via the familiar VIE structure — and deliver returns through traditional IPO exits.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Dear Readers,

With the current turmoil on the US stock market due to escalating tariff wars and rising recession fears, some investors are looking to actively diversifying their portfolios — shifting focus toward European and Asian assets.

However, Asia’s VC market struggled to gain momentum in February, showing even weaker performance than Europe and the US. Overall, tech PE/VC investments in APAC dropped by 15%, with China down 17% and India experiencing a 23% decline.

Despite the broader slowdown, AI and space tech stood out as growth sectors across the region. AI startups in India (Krutrim), Australia (Harrison.ai), and China (LillibAI) raised large funding rounds. Meanwhile, space tech saw major deals in Japan (Ark Edge Space) and China (RSpace), alongside new seed-stage startups emerging in India.

Compute and robotics continued to look promising, but China completely dominated—capturing ~99% of all APAC investments in February. Alongside traditional semiconductor PE rounds, next-gen quantum computing took off, with startups like CAS ColdAtom (neutral atom quantum computing) and Chinaprosp Quantum (diamond-based quantum sensors) securing major funding.

Another big trend: humanoid robots. Alibaba, Baidu, IDG, and CICC are backing the next wave of robotics leaders, betting on startups like Neolix and Galaxea to shape the future of automation.

Despite all this innovation, exits remain the biggest bottleneck for Asian VCs, especially in China. In February, China had just one small deep tech IPO and zero M&A deals. Among APAC markets, India is currently the most liquid, with four deep tech exits, but still trails far behind the US (31 exits) and Europe (15 exits).

In my next piece, I’ll take a closer look at the exit challenge, exploring what’s holding back liquidity in Asian deep tech and where the ecosystem goes from here. Stay tuned!

📃 Must Reads

💰 🌏 Deep Tech Funding Map in Asia (by DeepTech Asia)

If you're looking for a high-level overview of deep tech developments across Asia, our latest Deep Tech Funding Map is the perfect starting point!

We’ve analyzed total PE/VC funding across APAC, breaking it down by major deep tech segments—including AI, compute, robotics, climate, and space. China is certainly the biggest ecosystem in all these segments. However, while China remains challenging for some investors due to geopolitical reasons, identifying the second and the third largest players in each category becomes crucial, but also less obvious.

See the full version here.

🇨🇳 👾 Manus: China’s Latest AI Sensation (by ChinaTalk)

Absolutely recommend the excellent piece by ChinaTalk’s Lily Ottinger and Jordan Schneider about another Chinese AI phenomenon — general AI agent Manus. It is capable of autonomously performing complex tasks such as rebooking flights and managing data. Since 2023, Monica, the company behind Manus, has evolved from an AI-powered browser plugin into a comprehensive AI agent, securing investments from prominent funds like Tencent and Sequoia Capital China. Monica rejected a $30 million acquisition offer from ByteDance in early 2024, opting instead to raise additional funding, elevating its valuation to nearly $100 million.

More details here.

🤑 Deals (Feb 2025)

🇨🇳 ⚠️ China’s Investment Slowdown Continues. In February, China experienced a weak VC/PE funding environment, with total investments reaching $3.8B — lagging behind both Europe ($4.7B) and the US ($11.7B). Across the broader APAC region, tech companies raised $6.2B, marking a 15% decline compared to January.

🏯 📉 Investment Dip Across Asia. Beyond China, all major APAC markets saw a decline in funding compared to January. Tech investments in India fell by 23%, while Singapore experienced a sharp 78% drop. The most significant pullback was in climate tech and compute sectors, reflecting shifting investor priorities and a more cautious funding environment.

India is catching up with China in the AI race, Indian billionaire is putting more money into India’s Open AI contender, Australian AI healthcare startup raises a massive Series C round

🌏 Global scale of VC funding (Feb 2025):

1. 🇺🇸 USA — $3.4B, with $305M from Together AI and $300M from Harvey

2. APAC — $1.3B, inc. 🇨🇳 China ($816M), 🇮🇳 India ($304M) and 🇦🇺 Australia ($131M)

3. 🇪🇺 Europe — $737M, with $100M from DreamSecurity and $75M from Luminance

Highlights:

🇮🇳 $230M round from Bhavish Aggarwal: Krutrim, leading AI-focused cloud provider in India, developing their own open source AI models. The startup is founded by Aggarwal himself and is positioned as the first AI unicorn in India. The round was financed through Aggarwal’s family office

🇦🇺 $112M Series C from Aware Super and Horizons Ventures: Harrison.ai, a clinician-led healthcare AI company using deep learning for IVF patients.

🇨🇳 $42M Series A+ from Shunwei Capital: LiblibAI, an AI-powered productivity tool and content creation and sharing platform for the content creative industry.

🇮🇳 $16M Series A from Vertex SEA & India: Spyne, a startup helping businesses and marketplaces create and upgrade high-quality product images and videos at scale.

🇨🇳 $14M Series B+ from Baidu: Infimind (极睿科技), an AI content generation engine for e-commerce with massive fashion data and computer vision algorithms

US-based AI developer cloud provider raises a large round, late stage compute rounds in Europe, quantum startup boom in China

🌏 Global scale of VC funding (Feb 2025):

1. 🇺🇸 USA — $973M, with $480M from Lambda Labs

2. APAC — $445M, inc. 🇨🇳 China ($444M)

3. 🇪🇺 Europe — $431M, with $170M from Quantum Machines, $165M from Eclairion

Highlights:

🇨🇳 Strategic round from Wuhan Quantum fund: CAS ColdAtom (中科酷原), a quantum spin-off from China Academy of Science developing cold atoms-based 100-qubit quantum computers and quantum precision measurement and control tools.

🇨🇳 Strategic round from HKIC: StarFive (赛昉科技), developing CPU IP, NoC IP, SoC, and development boards based on RISC V instruction sets.

🇨🇳 Series A: Chinaprosp Quantum (国盛量子), a developer of diamond-based quantum sensors for domestic and global industrial use cases.

🇨🇳 Series A+ from Beijing Industrial Fund: QBoson (玻色量子), a startup developing 550-qubit optical quantum computing equipment.

🇰🇷 $1M round from CRIT Ventures: FuriosaAI, a startup designing and developing data center accelerators for the most advanced AI models and applications.

US mega rounds, Baidu and Alibaba making early-stage bets on future humanoid robot leaders in China, Lightspeed investing in a Chinese AI agent startup

🌏 Global scale of VC funding (Feb 2025):

1. 🇺🇸 USA — $1.1B, with $600M from Saronic and $350M from Apptronik

2. APAC — $377M, inc. 🇨🇳 China ($375M)

3. 🇪🇺 Europe — $61M, with $44M from Nomagic

Highlights:

🇨🇳 $140M Series C+ from CICC Capital: Neolix (新石器无人车), a startup focusing on R&D and manufacturing of unmanned cargo vehicles with the L4 autonomy level.

🇨🇳 $42M Series A from IDG, Ant, Baidu and Hillhouse: Galaxea (星海图), a company building humanoid robots and advanced robotic arms.

🇨🇳 $42M Series A+ from Alibaba: CoreNetic (源络科技), a robot developer, focusing on the R&D and application of robot embodied intelligence technology.

🇨🇳 $42M Series pre-A from Legend Capital, Lightspeed: X Square Robot (自变量机器人), a startup developing general AI agents based on LLMs for robotic applications.

🇮🇳 $2M Seed round: Anscer Robotics, a startup building autonomous cleaning robots for nuclear waste.

Climate investments in Asia keep declining while Europe is leading the charge in this space, Chinese nuclear fusion project secures massive funding from CNNC

🌏 Global scale of VC funding (Feb 2025):

1. 🇺🇸 USA — $878M, with $270M from TransGrid Energy

2. 🇪🇺 Europe — $873M, with $615M from Enilive

3. APAC — $470M, inc. 🇨🇳 China ($418M), 🇮🇳 India ($17M), 🇯🇵 Japan ($15M)

Highlights:

🇨🇳 $245M round from CNNC and Zheneng Electric: Fusion Energy (中国聚变能源), a controlled nuclear fusion technology company under CNNC dedicated to pooling resources to accelerate nuclear fusion R&D in China.

🇨🇳 $3M Series A from CAS Star: Yili Technology (宜锂科技), a developer of solid state battery cathode material for power batteries, consumer, and drone batteries.

🇯🇵 $12M Series B from Keio Innovation Initiative: Aeterlink, a manufacturing company that develops wireless power transfer technology for wireless devices.

🇮🇳 $5M Series A from Unicorn India Ventures: Probus Smart Things, a company providing an IoT solution for electricity distribution utilities.

🇸🇬 $2M Seed: HYDGEN, a clean tech startup empowering empowers industries to produce their own green hydrogen on-site.

Japanese microsatellite manufacturer secured Series B, more seed startups from India, Chinese space manufacturing startup raising a massive Series A

🌏 Global scale of VC funding (Feb 2025):

1. 🇺🇸 USA — $195M, with $100M from K2 Space

2. APAC — $130M, inc. 🇨🇳 China ($76M), 🇯🇵 Japan ($52M), 🇮🇳 India ($3M)

3. 🇪🇺 Europe — $40M, with $21M from EnduroSat

Highlights:

🇯🇵 $52M Series B from Incubate Fund: Ark Edge Space, a manufacturing firm that offers microsatellite design, development, production, and operational services.

🇨🇳 $42M Series A+: RSpace (九天行歌), producer of room temperature and low temperature propellant tanks for liquid fuel rockets.

🇨🇳 $14M Series A+: Hongqing Technology (鸿擎科技), a satellite networking solution provider for low-orbit broadband Internet communication.

🇮🇳 $1M Seed round: Sisir Radar, a space and defense technology startup specializing in advanced Synthetic Aperture Radar (SAR) systems.

🇮🇳 $1M Seed round: Astrogate Labs, a spacetech startup specializing in high-speed laser communication systems for small satellites.

Exits remain the major challenge for the Asian VC ecosystem, India is leading the way, but still far behind Europe and the US

🌏 Global scale of VC-backed exits (Feb 2024):

1. 🇺🇸 USA — 100 exits (31 deep tech exits)

2. 🇪🇺 Europe — 58 exits (15 deep tech exits)

3. APAC — 47 exits (20 deep tech: 🇨🇳 China - 14, 🇮🇳 India - 4, 🇦🇺 New Zealand - 1, 🇦🇪 UAE - 1).

Highlights:

🇮🇳 $2.2B M&A: Ayana Renewable Power, an energy generation company that focuses on the development of the renewable energy sector, was acquired by Oil and Natural Gas Corporation. Exit for EverSource, NIIF and British International Investment.

🇨🇳 IPO at a $179M market cap: Changyou Technology (常友科技) , a wind power composite materials startup. Exit for EcoFund and China VC.

🇮🇳 M&A: Clari5, an AI-powered financial crime detection and risk management platform was acquired by Perfios, an Indian B2B SaaS Fintech company. Price is not disclosed.

🇦🇪 M&A: Tazzweed, a digital procurement and outsourcing platform for sales teams was acquired by Neu Ocean Technologies, a UK-based startup. Price is not disclosed.

🇮🇳 M&A: Grinntech, a designer of lithium battery packs was acquired by Yuma Energy, a battery-as-a-service JV of a Canadian auto component maker Magna International and Indian mobility startup Yulu. Exit for angel investors.

🇦🇺 M&A: Robotics Plus, a New Zealand-based robotic apple packer was acquired by the Japanese corporation Yamaha Motor.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

🇨🇳 China leads deep tech in Asia, dominating robotics (93% of APAC funding) and space tech (82%), while also pushing AI infrastructure and autonomous vehicles. In 2024, it outpaced the US in space tech funding ($2.7B vs. $2.2B).

🇸🇬 Singapore is Asia’s largest AI hub (outside China), attracting global startups (like MiniMax) and corporations (like GDS) thanks to its pro-business policies and great investment climate.

🇯🇵 Japan remains strong in robotics and some other deep tech areas but slow growth and corporate-dominated market make scaling startups tough, despite AI talent inflow.

🇦🇺 Australia is a rising space tech and climate leader, benefiting a lot from its rich natural resources, solid scientific base and strong US ties.

🇮🇳 India and 🇰🇷 South Korea are emerging. India is growing in climate tech and robotics, while South Korea’s semiconductor strength fuels AI and compute tech.

With the ongoing US-China rivalry and the shifting global supply chains, it's clear that Asia is emerging as a major hub for tech innovation. But understanding the true scale can be tricky without digging into the exact numbers.

In 2024, the US was still the heavyweight champion in venture funding, pulling in over 50% of global VC and tech PE investments. The US was the only major market that showed growth in 2024 (36% YoY), while others took a hit—Europe dropped 6%, and China and India both slid by 12%. The reasons? A business-friendly environment, a massive unified market and liquid capital markets.

Asia is no small player either, hitting $118B in VC and tech PE funding in 2024—75% bigger than Europe’s but still only about half the size of the US. China remains the region’s biggest driver, accounting for 66% of Asia’s tech funding, though its share slipped slightly from 69% in 2023.

What about deep tech?

The definition of deep tech various significantly across different sources, so we’ll not attempt to come up with a figure that would exactly define the total amount of all deep tech investments. Instead, we will focus on specific industries that we assume to be “deep tech”.

Let’s start with the area is now probably “the least deep tech” — Artificial Intelligence. Yes, it may be considered deep tech five years ago, but now almost every company is an AI company.

Nevertheless, AI is important in our context, because it dictates the trend for many other “truly deep tech” segments.

In 2024, it seemed that every deal was an AI deal. In the US, 37% of all investments were somehow “AI-related”, compared to 25% in 2023. However, in other markets it wasn’t really the case. In Europe and APAC it was only ~15%, 17% in China, 18% in Singapore and only 3% in India. US is clearly dictating the talk of the town!

🇨🇳 China is clearly the largest source of AI funding in Asia with 76% of all Asian AI investments in 2024 coming from there. This included:

• Zhipu AI — the Chinese AI champion Zhipu AI that raised $1.3B in three rounds (Series C, Series D and a strategic round) in one year,

• Moonshot AI with its Kimi AI assistant that raised $1.2B in Series A and B, also in one year,

• IM Motors (智己汽车), the JV between SAIC Motors, Alibaba and Zhangjiang Hi-Tech manufacturing AI-powered EVs that raised a massive $1.1B Series B round.

🇸🇬 🇯🇵 🇰🇷 Singapore, Japan and South Korea also saw some prominent AI rounds:

• MiniMax — Singaporean developer of AI video generation app Hailuo that raised $600M Series B from Alibaba and Hongshan (ex-Sequoia China),

• Sakana AI — Japanese LLM champion that in one year raised $244M in seed and Series A funding from Lux, Khosla and NEA,

• DeepX — Korean AI chip and SDK developer which closed a $163M Series C round from local PE and VC funds.

AI is fueling an explosion in demand for new computing technologies, cranking up the tech rivalry between the US and China. The US is way ahead, pouring nearly $21B into semiconductors and next-gen compute tech (quantum, photonics), keeping its edge in the race for AI dominance.

🇨🇳 China is up against tough sanctions and restrictions, cutting off access to top-tier tech from TSMC, ASML, and other global leaders. In response, it's going all-in on building its own semiconductor industry. In 2024, China pumped over $15B into new compute startups—a whopping 78% of all Asian investments in this space. But here’s the catch: most of these deals are tech PE rounds, not traditional VC rounds. For example:

• BEHC (北电集成) — Chinese producer of digital-analog chips, MCUs and display driver chips ($2.8B round),

• ChangXin Memory (长鑫存储) — Chinese dynamic random memory (DRAM) producer ($1.5B round),

• Hefei Wanxin (合肥皖芯) — Chinese producer of 56-nm and 28-nm display chips ($1.3B round).

🇸🇬 Notably, Singapore is actively positioning itself as a regional data center hub, attracting a lot of large global players:

• Chinese GDS data center operator that established its international headquarters in Singapore and raised $1.6B from Coatue and Hillhouse in 2024

• ST Telemedia Global Data Center, a consortium of Singtel and KKR, raised $1.75B.

🇰🇷 Also, South Korea is becoming strong in the AI accelerator area with the news rounds of DeepX ($163M) and Rebellions ($140M).

AI breakthroughs—especially in large language models (LLMs)—are kicking off a robotics boom worldwide. Humanoid robots and self-driving cars, once straight out of Westworld and sci-fi movies, are inching closer to reality. The US is leading the charge, with monster funding rounds like Waymo’s $5.6B Series C for autonomous vehicles and Figure’s $675M Series B to build humanoid robots. The future we used to dream about? It’s finally coming to life.

🇨🇳 In Asia, China is an absolute dominant force in robotics, accounting for whooping 93% of all Asian investments. The biggest investments include:

• YinWang Intelligence (引望智能) — a startup under Huawei developing software systems for autonomous vehicles that raised $3.2B round from Changan’s Avatr and another Chinese automaker SOKON,

• The AV subsidiary of the ride-hailing company DiDi raised a $271M Series C,

• Windrose (苇渡科技) — developer of autonomous EV trucks raised $200M in its Series B+ and pre-IPO round from HSBC and GSR Ventures.

🇯🇵 Japan has also been historically strong in robotics. In 2024, its industrial automation startup Preferred Networks raised a $121M round from SBI. But generally the investments in Japanese robotics companies are declining.

🇰🇷 Another emerging leader is South Korea which has a few promising startups such as LIDAR developer Bitsensing (raised $25M Series B), medical robotics startup AIRS Medical (raised $20M Series C) and self-driving software developer RideFlux (raised $19M Series B).

In 2024, the gap in climate funding between the US and Europe widened big time—from $2B in 2023 to $10B. The surge was driven by massive PE deals in US renewables, like Avangrid Renewables ($2.5B) and Pine Gate Renewables ($938M), along with huge VC rounds for nuclear fusion startup Pacific Fusion ($900M Series A) and carbon capture company Twelve ($600M). But things could shift in 2025, especially with Trump’s second term potentially shaking up US climate policy.

In Asia, climate funding took a noticeable dip compared to 2023, with significant decreases in China (-23%), India (-21%), and Australia (-68%).

🇨🇳 Despite this downturn, China still accounts for a whopping 74% of all climate investments in the region. The biggest investments are pouring into corporate and venture rounds for up-and-coming EV makers like Avatr ($1.6B Series C) and BAIC Blue Park ($1.1B). Renewable energy players are also attracting massive funding, with China Huaneng securing $2.7B, CNNP Rich Energy raising $1.1B, and China Hydrogen Technology pulling in $910M. These numbers highlight the growing momentum in green tech and clean energy.

🇮🇳 India is the second largest energy market in Asia with $3.3B funding in 2024, however, most of it are PE rounds of renewable companies, such as SAEL ($1B), the JV of Total Energies and Adani Green Energy ($440M+) and Fourth Partner Energy ($275M). On the venture side, here are some of the biggest rounds in India:

• Battery Smart — EV battery swapping platform that raised a $110M Series B,

• Ather Energy — provider of electric scooters and fast EV charging ($82M),

• Lohum — developer of lithium-ion batteries and recycling tech ($77M).

🇨🇳 🇺🇸 The space race is heating up, and the US and China are throwing more money than ever into private space companies. In 2023, US space tech startups pulled in $3.1B in venture funding—nearly 3x more than China. But 2024 flipped the script —for the first time, China’s private space sector outpaced the US, raising $2.7B vs. $2.2B. Many of these Chinese startups keep a low profile, but we’ve got the scoop on some of the biggest funding rounds below:

• Yuanxin Satellite (垣信卫星) — Chinese competitor of Starlink operating satellite constellation on a low Earth orbit and providing communication, navigation and remote sensing services for the commercial sector raised a $938M Series A',

• Space Pioneer — manufacturer of reusable rockets raised a $210M Series C+,

• FOSST — operator of a communication satellite constellation raised $140M.

Despite the fact that China accounts for 82% of all space tech investments in APAC, there were other sizeable space investments in other parts of the region, first of all, Australia, UAE and Japan:

• 🇦🇪 Orbitworks — Abu Dhabi-based integrator of remote sensing satellites, a JV between US-based Loft Orbital and UAE-based Marlan Space, raised $100M,

• 🇦🇺 Fleet Space Technologies — Australian space-enabled mineral exploration startup raised a $100M Series D,

• 🇯🇵 Japanese Synspective providing high-resolution satellite data and remote monitoring services for government agencies raised a $44M Series C.

China is still the largest Asian ecosystem

Drawing a clear line between “deep tech” and “non-deep tech” is tricky, however, it’s safe to say that China is by far the largest player in the deep tech ecosystem in Asia. Although China is losing its share in the overall VC/PE market (from 69% in 2023 to 66% in 2024), it still dominates every deep tech vertical, especially robotics (93% of all APAC investments) and space tech (82%).

🇨🇳 China still lags behind the US in funding across many sectors, especially AI ($81B vs. $13B). But that doesn’t mean Chinese AI models can’t keep up. The recent hype around DeepSeek, a Chinese open-source LLM developer, proves they’re very much in the game. If you want a deep dive, ChinaTalk’s latest article is worth a read. That said, the funding gap does point to a key difference— LLMs in China aren’t fueling the ecosystem of AI applications as effectively as they are in the US.

That said, the Chinese government is all in on AI—pushing for breakthroughs in AGI, AI hardware, autonomous vehicles, and robotics. In these areas, China’s VC/PE funding is only 30-40% behind the US, a much smaller gap compared to the 6x difference in overall AI funding.

One area where China actually pulled ahead in 2024 was Space tech. Chinese startups raked in $2.7B vs. $2.2B in the US, signaling that China is doubling down on private space ventures to go toe-to-toe with the US in the new space race.

The second largest ones

While China is an obvious technological leader in Asia, it’s hard to pick the second strongest player.

🇸🇬 Singapore is riding the AI wave thanks to its business-friendly policies and strong investment climate. More and more regional startups are setting up shop there, and Chinese companies are increasingly using Singapore as a launchpad for global expansion. With a highly skilled workforce and a solid semiconductor industry, Singapore has cemented itself as Asia’s biggest AI and advanced computing hub outside of Mainland China in 2024.

🇯🇵 Japan was Asia’s tech powerhouse back in the 1980s, and while it’s no longer on top, its strong robotics industry has kept it in second place after China. Japan is also working hard to attract top AI talent, leading to startups like Sakana AI setting up shop in Tokyo. But despite these efforts, Japan still struggles with sluggish economic growth, and its VC scene is dominated by big corporations, making it tough for smaller startups to break through.

🇦🇺 Australia is shaping up to be a major tech hub in APAC. In 2024, it ranked as Asia’s second-largest space tech VC market, thanks to its strong scientific base and close ties with the US. On top of that, its rich natural resources make it a huge player in energy and climate investments.

🇮🇳 India has huge potential thanks to its massive population and market size, but when it comes to deep tech, it’s still playing catch-up with other developed Asian economies. That said, there are a few bright spots—India has solid momentum in energy and climate tech, and its robotics and drone industry is starting to take off, though it’s still in the early stages.

🇰🇷 South Korea deserves a shoutout for its massive semiconductor industry, home to giants like Samsung and SK Hynix. This strong foundation has made it Asia’s second-biggest market for new compute technologies, helping the country build serious momentum in AI applications and robotics.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Dear Readers,

If you were to name one word that came to your mind in January in relation to Asia, what would it be? — “DeepSeek”. It’s been a while since Asia, and specifically China, have been given that much attention globally, which is a good thing for both the West and the East.

For the West it is a “wake-up call” (quoting Trump) and the incentive to increase its efficiency and competitiveness, while for the East it’s an aspiring example that Asian cutting-edge startups can compete with their American and European opponents not only on price, but on quality and performance as well.

Below I share some other examples of promising deep tech startups coming out of Asia, apart from DeepSeek. In January, China led the charge in robotics with several big rounds of medical robotics companies led by global investors like EQT, Vertex, Medtronic and Saudi Aramco’s Prosperity7. Chinese startups in this area attracted $446M of VC money, compared to $216M in Europe and $169M in the US.

In space tech, Japan and India are looking stronger and stronger with new investments in satellite and space travel startups. Meanwhile, Singapore is further strengthening its status as a data center hub in Asia with the new mega round of Digital Edge collocation operator.

Having said this, exits still remain one of the biggest challenges for Asia, with only 41 VC-backed exits in January in the whole APAC region (including China), lagging behind Europe (53 exits) and by far the US (129 exits).

📃 Must Reads

🐳 💰 Would you invest in DeepSeek if you had a chance? (by DeepTech Asia)

On the hype around DeepSeek, we’ve tried to differentiate and present a thought experiment on whether DeepSeek could be a good investment opportunity in case it goes out raising money. Still a very hypothetical scenario. The company is funded by its parent fund and the founder is not focused on commercialisation in the near term.

But I was really surprised to hear that Dylan Patel from SemiAnalysis also mentioned in his interview to Lex Friedman that DeepSeek is burning cash and may need funding soon. Also there is a rumour that DeepSeek actually tested the fundraising market in China in May last year, but eventually decided to keep bootstrapping.

Anyway, here you can find my “mini investment memo” on DeepSeek. I hope you’ll enjoy it! And below are the results of the survey:

🇨🇳 👾 Chinese AI companies that could match DeepSeek’s impact (by the Guardian)

A brief snapshot of other leading LLM companies from China that are competing with DeepSeek on their domestic market and could be joining the global race soon. The Guardian listed a few models developed by large corporates like Alibaba, Bytedance, Tencent, as well as, leading startups like Zhipu AI and Moonshot AI.

More details here.

🤑 Deals (Jan 2025)

🐍🏮 Lunar New Year in Asia. DeepSeek made quite a stir during the Lunar New Year when many investors and entrepreneurs in Asia enjoy holidays and time with their families. That is why the January period was pretty quite in Asia in terms of VC investments. For example, China only saw $4.6B investments which was lower than in Europe ($5.7B) and by far than the US ($18.7B). In total, tech companies from APAC attracted $7.3B in VC/PE funding in January.

🇮🇳 India was the second largest VC location in Asia. In January Indian startups attracted $1.3B. Singaporean startups also were quite active, securing $745M of VC/PE funding, while the biggest part was contributed by the PE round of a data center company — Digital Edge DC.

US is spearheading the AI funding, China doubles down on AI applications and Hedosophia backing an Aussie AI media startup

🌏 Global scale of VC funding (Jan 2025):

1. 🇺🇸 USA — $3.6B, with $1B from Anthropic

2. APAC — $1.2B, inc. 🇨🇳 China ($1B), 🇮🇳 India ($48M) and 🇯🇵 Japan ($34M)

3. 🇪🇺 Europe — $700M, with $180M from Synthesia

Highlights:

🇨🇳 $91M Series E: Insilico (英矽智能), utilising generative AI for drug discovery in the areas, such as cancer, fibrosis, immunity, central nervous system and aging-related diseases.

🇨🇳 $70M Series D from Legend Capital: Deepwise (深睿医疗), AI-powered solution for radiologists focused on diagnosis accuracy multiple applications, such as mammography, orthopaedics, neurosurgery and others.

🇨🇳 $42M Series Pre-A: AI² Robotics (智平方), developing LLMs for robotics applications with a focus on warehouses and households.

🇦🇺 $25M Series B from Hedosophia: Qsic, an AI driven experience platform focusing on in-store audio for retail media.

🇮🇳 $25M Series C: SuperOps.ai, an AI-powered professional services automation (PSA) and remote management monitoring (RMM) for MSPs.

US keeps plowing cash into more data centers, Vertex investing in a Chinese auto chip manufacturer and a new Japanese quantum startup secured seed funding.

🌏 Global scale of VC funding (Jan 2025):

1. 🇺🇸 USA — $5.1B, inc. the $5B PE round of Aligned

2. APAC — $1.9B, inc. 🇨🇳 China ($1.3B), 🇸🇬 Singapore ($640M), 🇦🇺 Australia ($37M)

3. 🇪🇺 Europe — $224M

Highlights:

🇸🇬 $640M PE round: Digital Edge DC, a Stonepeak-backed developer of carrier-neutral data centers and related digital infrastructure assets.

🇨🇳 $40M+ Series A from Lightspeed China: Infrawaves (基流科技) , providing hardware and software solutions for GPU cluster interconnection, including high-speed switches and management and communication frameworks.

🇨🇳 $42M Series A from Vertex: MetaSilicon (元视芯), designing integrated circuits for automotive applications.

🇦🇺 $20M Series A extension: Quantum Brilliance, a developer of quantum computers that operate at room temperature using synthetic diamonds.

🇯🇵 $4M Seed from Global Brain: OptQC, developing optical quantum computers based on the Furusawa-Endo Laboratory research.

China leading the charge in robotics globally, rise of medical robotics startups and more global investors betting on Chinese robotic companies.

🌏 Global scale of VC funding (Jan 2025):

1. APAC — $453M, inc. 🇨🇳 China ($446M)

2. 🇪🇺 Europe — $216M, with $124M from Neura Robotics

3. 🇺🇸 USA — $169M

Highlights:

🇨🇳 $70M Series C+ from EQT, Qiming, Alpha JWC: Cornerstone Robotics (康诺思腾), developer of soft tissue and other major specialized surgical robots.

🇨🇳 $42M Series E+ from Prosperity7: Fourier (傅利叶), developing humanoid robots and robotics solutions for rehabilitation applications.

🇨🇳 $14M Series B from Vertex: AI Force Tech (中科原动力), provider of robotic, automation and big data solutions for agriculture.

🇨🇳 $7M Series C from Medtronic: Rossum Robot (罗森博特), focusing on medical robotics, such as real-time intraoperative 3D navigation, assisted fracture reduction operation, automatic surgical planning.

🇯🇵 $3M Seed from JAFCO: ARCS, developing assisted reproductive medical devices using robotics and artificial intelligence technology.

Slowdown in climate investments in Asia, Temasek backing a mega Series A of a Chinese battery material startup and Walden betting on autonomous industrial EVs in India

🌏 Global scale of VC funding (Jan 2025):

1. 🇺🇸 USA — $1.8B, with $1B from Origis Energy

2. 🇪🇺 Europe — $1.1B, with $400M+ from Green Flexibility

3. APAC — $605M, inc. 🇨🇳 China ($486M), 🇮🇳 India ($62M), 🇰🇷 South Korea ($21M)

Highlights:

🇨🇳 $140M Series A+ from Temasek and TPC Group: Nanopore (纳力新材料), developing new composite current collector materials for lithium batteries.

🇨🇳 $42M Series B from Hongshan (ex-Sequoia China): PuXiJing (朴烯晶), developing high-end polymer materials for lithium battery separators.

🇮🇳 $20M Series B from Walden Catalyst: Ati Motors, an all-electric autonomous industrial vehicle maker.

🇰🇷 $16M Convertible Note from STIC Investment: H2, an advanced energy storage company specialized in vanadium redox flow battery.

🇳🇿 $11M Series A from Blackbird Ventures: Kwetta, an EV charging startup enabling deep integration with power grids.

US is supercharging Space Tech, Toyota’s VC fund doubling down on Interstellar Technologies and more seed deals from India

🌏 Global scale of VC funding (Jan 2025):

1. 🇺🇸 USA — $520M, with $260M from Stoke Space and $170M from Loft Orbital

2. APAC — $129M, inc. 🇨🇳 China ($60M), 🇯🇵 Japan ($44M), 🇮🇳 India ($23M)

3. 🇪🇺 Europe — $75M, with $41M from Xcalibur

Highlights:

🇯🇵 $44M Series F from Toyota’s Woven: Interstellar Technologies, developing cost-effective and convenient space transportation services.

🇨🇳 $42M Series A+ from Kunlun Capital: Hongqing Technology (鸿擎科技), a satellite networking solution provider, focusing on low-orbit broadband Internet communication satellite applications.

🇨🇳 $14M Series A from Shenzhen Capital: Azimuth Technology (方位角), providing high-precision spatiotemporal and indoor / outdoor integrated positioning and navigation services.

🇮🇳 $1.7M Pre-Seed: Catalyx, a space travel startup.

🇮🇳 $1.5M Pre-Seed: OrbitAID Aerospace is developing satellite life extension solutions enabling longer mission duration, reduced space debris generation.

Asian exits lagging behind, British fund Eight Roads on track to exit from its Chinese medical AI investment and Aussie energy tech startup getting acquired

🌏 Global scale of VC-backed exits (Jan 2024):

1. 🇺🇸 USA — 129 exits (42 deep tech exits)

2. 🇪🇺 Europe — 53 exits (14 deep tech exits)

3. APAC — 41 exits (13 deep tech: 🇨🇳 China - 11, 🇦🇺 Australia - 1, 🇸🇬 Singapore - 1).

Highlights:

🇨🇳 IPO at a $457M market cap: BrainAu (脑动极光), AI startup developing medical-grade digital therapy products for cognitive disorders. Exit for Eight Roads and Northern Light Capital.

🇨🇳 IPO at a $429M market cap: Hyper Strong (海博思创), a provider of large-scale energy storage and EV battery systems. Exit for IDG Capital and Qiming Ventures.

🇦🇺 M&A: Everty, providing vehicle charging station management services for residential, business, and public locations; acquired by Australian renewable energy producer AGL Energy. Exit for Artesian VC, EnergyLab and Sydney Angels.

🇨🇳 $304M acquisition: EverProTech (长芯盛), developing and producing photonic chips, optoelectronic modules, high-speed active optical cable (AOC) interconnects; acquired by a Chinese fiber optics manufacturer Broadex Technologies.

🇨🇳 IPO at a $194M market cap: Saimo Technology (赛目科技), a digital twin startup focusing on simulation tests for intelligent connected vehicles. Exit for Huawei Hubble and CITIC Securities.

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Okay, now it’s my turn to write about DeepSeek. It’s hard to find someone who doesn’t write about this Chinese AI company these days.

No surprise.

DeepSeek erased trillions of dollars from the American stock market, NVIDIA was down 21% (their biggest dip ever), and the global media are panicking about the threat from Chinese tech which they claimed to be dead just a year ago.

There are already a lot of good articles about DeepSeek, but I haven’t seen much about the investment prospects related to the DeepSeek itself.

So, I decided to dive into it myself.

Just as a thought experiment, let’s assume that DeepSeek is raising a private round and you as an investor receive some allocation in this round. Would you invest?

Sure, many smart investors would say that DeepSeek is not a startup, it’s a subsidiary of a hedge fund which is fully funded by its parent company and doesn’t require any capital. However, DeepSeek is still at an early stage and for its further product development, commercialisation and expansion the capital of its parent company High-Flyer might not be enough.

In the past we’ve seen some open-source startups emerging as spin outs from other corporations and raising money from Tier1 investors, e.g.:

• Confluent, a spin-out from LinkedIn, raised $455M from Benchmark, Sequoia, Coatue, and Index.

• Clickhouse, a Yandex spin-off, raised $300M from Bessemer, Index and Coatue.

• Heptio, a spin-out from Google, raised $33M from Accel and Lightspeed and got acquired by VMware for $550M.

Perhaps DeepSeek could also follow the same path and become “investable”. Below I’ll share my view on the key criteria for such a theoretical investment decision.

Investment structure — 6/10

First of all, if you are a non-Chinese investor managing assets in any currency other than RMB, investing directly in DeepSeek might not be feasible. Considering the attention company is now getting from the Chinese government, it may be designated as a national project in China, which typically restricts foreign investment unless you're a strategic state fund like Mubadala or Temasek.

However, foreign investors still could invest via an special purpose vehicle (SPV) or even a feeder fund managed by one of those lucky private funds that managed to get allocation in the company.

In most cases, these investment vehicles may even have a friendly jurisdiction for foreign investors, e.g. Caymans, Bermuda or even Delaware, which is usually the case for many other late stage Chinese unicorns (Bytedance, Red Note, Shein etc).

Valuation is hard to guess, but we may compare DeepSeek now with Open AI in early 2023 when it just released ChatGPT and raised $10B from Microsoft. Back then, Open AI also didn’t have any meaningful revenue and was valued at $29B.

You may apply a 50% China discount (e.g. P/S ratio for Alibaba is 50% lower than for Amazon), which would bring us to around $15B.

For the context, Zhipu AI, another Chinese open source LLM provider that is still behind Open AI in terms of performance, has recently raised a $412M round at $3B valuation.

Risks:

• Limited transparency — medium. Although non-Chinese investors may invest through structures in familiar jurisdictions, having additional layers between them and the company creates complexity. Limited transparency increases chances of fraudulent schemes. Additionally, there are also language and cultural barriers that should not be underestimated.

Technology — 10/10

DeepSeek has released several AI models over the past year, including DeepSeek LLM with a long-term scalability perspective, DeepSeek-VL for real-world vision and language understanding applications and DeepSeek-V2, a Mixture-of-Experts (MoE) model with highly efficient training and inference.

However, the biggest stir was caused by the recent two models released in January 2025, namely, DeepSeek-V3, a stronger version of V2 with 671B parameters and 37B activated for each token, and DeepSeek-R1, their first reasoning model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT).

V3 outperforms other open-source AI models and achieves performance comparable to the leading closed-source models, such as GPT-4o and Claude-3.5-Sonnet, while R1 outperforms OpenAI-o1 in most tasks, including reasoning, knowledge and math and is comparable with OpenAI-o1 in coding.

The major advantage of DeepSeek models is their efficiency. In its paper DeepSeek claims to have used only 2048 H800 GPUs for less than two months for its V3 model training, which is an equivalent of 2788k GPU hours. Assuming that the rental price of H800 is $2 per GPU hour, they estimated the compute costs on training only at $5.576M. It excluded any staff, infrastructure and energy costs.

For comparison, OpenAI spent $100M on training ChatGPT4, including $40M of the total compute cost and $18M in pure AI chip cost, according to this Stanford paper. Considering that staff cost in Hangzhou is considerably lower than in Silicon Valley, the total training cost for DeepSeek-V3 should be around $20-25M, which is 4-5x times lower than for the comparable ChatGPT4 model.

Despite its technological advancement, DeepSeek’s product is still at a very early stage. It doesn’t support the full multimodality and is inferior in terms of security compared to other leading AI models. However, those problems will be solved as the product matures, e.g. DeepSeek has recently announced its first multimodal model Janus-Pro.

Risks:

• Security — low. Security researchers keep finding a lot of vulnerabilities in DeepSeek models exposing prompts and internal data of users. It’s common for an early stage company like DeepSeek, they would need to invest more resources in fixing this, but it’s doable.

• Data privacy — low. A lot of users or potential customers are concerned about their data potentially accessed by the Chinese government. It is probably the case for individual users. However, any enterprise customers can deploy DeepSeek on the servers they trust, out of China. There was an interesting article by Interconnect explaining this.

Conclusion: DeepSeek has made an impressive progress in a limited amount of time with extremely constrained resources; technology is clearly the biggest advantage of the company.

Product and Business Model — 7/10

Although DeepSeek is promoting open source, we know a lot of successful commercial open source companies, such as Red Hat, MongoDB and Databricks, including several companies originated from China, e.g. PingCap.

Theoretically DeepSeek could also be one of them.

Closed-source horizontal LLM developers mainly monetise through their B2C product subscription, while some revenue portion comes through APIs. For example, Open AI generates 75% of its revenues through the ChatGPT subscriptions, while open source AI models like Llama keep their core product free and monetise mainly through APIs. At least for now.

Chinese LLM developers, vice versa, mainly monetise through APIs and MaaS (Model-as-a-Service) and on top of this have some B2C subscription revenue, which has been the case in China for many other software businesses. Some Chinese open source AI companies, such as Zhipu AI, provide some of their models for free, while keeping the other models proprietary.

Now, DeepSeek’s business model is similar to Llama’s one, it has two working products: 1) DeepSeek Chat which is free of charge, and 2) DeepSeek API.

While DeepSeek-V3 APIs are 18x cheaper than ChatGPT-4o and other Western AI models, their pricing is almost the same as for their Chinese AI developers. It means that now there will be a pricing war between LLMs globally, but first of all, in China.

Unit economics is another question. In his interview in July 2024, the DeepSeek’s founder Liang Wenfeng mentioned that the DeepSeek has “modest profit margin above its costs”. However, it’s not clear if he included the full cost in his calculations.

Open AI is losing money on its models even with the current pricing. Considering that DeepSeek’s price is 18x lower than for Open AI, but the total training cost is potentially only 4-5 times lower, DeepSeek should now have ~4x lower margins than Open AI. It’s definitely not sustainable in the long run, unless DeepSeek raises money.

Risks:

• Export controls — high. Trump is already considering tightening restrictions for chip exports to prevent Chinese companies like DeepSeek from getting enough GPUs. Of course, DeepSeek would still get them in one way or another, but the cost could rise significantly, which would make the cost advantage less critical.

• Product development — low. Yes, DeepSeek still needs to develop a lot of features, such as multimodality, as well as optimise the existing models. However, with such a strong team it’s a matter of time.

Conclusion: the cost and price advantages are clearly impressive, but it’s still questionable if the company manages to maintain them with further tightening of export controls and the growth of the business.

Market and Competition — 4/10

China’s AI market size is projected to reach $100B in 2025, according to Qianzhan Research Institute and it’s going to be a bloodbath among corporates (Bytedance, Alibaba, Tencent etc) and large AI startups (Zhipu AI, 01.ai, Moonshot AI, potentially DeepSeek).

However, the biggest advantage of DeepSeek is that it’s open-source. Their biggest opportunity is the global market which is counted in trillions of dollars.

Some may say that no American or European companies would ever use a Chinese LLM because of censorship and data privacy. I wouldn’t agree with this statement, and I bet that Western companies are already deploying DeepSeek. As we mentioned above, as long they deploy DeepSeek on the servers they trust, they are relatively safe.

Risks:

• Commoditisation — high. DeepSeek is pushing the prices for AI models down, thus decreasing margins for all LLM developers across the industry. In the short term, it may give advantage to DeepSeek, but in the long run it decreases the total addressable market.

• Geopolitical barriers — high. Western countries will most likely prohibit sensitive industries (defence, finance etc) from using DeepSeek as infrastructure. There still may be enough non-sensitive segments on the Western markets that DeepSeek could penetrate. However, if China’s government decides to “participate more actively in DeepSeek’s development” and invest in it through one of its state funds, it can potentially jeopardise relationships between DeepSeek and its potential Western customers.

• Technology diffusion — medium. Large companies like Alibaba and Bytedance may just adopt DeepSeek’s training method because it’s open source and eliminate the first-mover advantage.

• Global competition — medium. DeepSeek has cost advantage against the US competitors. Global companies are not as aggressive as Chinese companies in terms of pricing. However, they will still adopt DeepSeek’s technology sooner or later if it proves to be a more efficient way to train AI models.

• Challenges with global expansion — low. Previously there were not that many Chinese (or Asian) companies that succeeded in expanding globally, but the trend is changing with Huawei, BYD, CATL, Bytedance and others.

• Government crackdown in China — medium. This risk has shattered the investment sentiments in China a lot over the past few years. AI is one of the national priorities for China, so this sector is pretty safe from any scrutiny from the government. However, at some point anything may change, so it is important to follow the state agenda closely.

Conclusion: the market is huge, but there are a lot of market risks associated with DeepSeek being a Chinese company expanding globally, as well as industrial risks for any LLM developers.

Team — 6/10

DeepSeek has around 140 employees, mostly scientists and developers who are based in China (Hangzhou, Shanghai, Beijing). But there are also some people in the US (California, Texas) who mainly cover sales and partnerships.

Liang Wenfeng (梁文锋), the founder, is very low profile. He was born in 1985, graduated from from Zhejiang University in 2008 with a major in Information and Electronic Engineering. In 2015, he created High-Flyer (幻方量化), one of the largest quant hedge funds in China ($7B AUM). In 2016, they started to experiment with AI and fully adopted AI for all their trading operations in 2017. In 2019, Liang spearheaded the team focused on training an advanced AI platform Firefly I (萤火一号). Since 2019, they’ve invested $140M in the Firefly project before it became DeepSeek in 2023.

Some people admit that Liang is not motivated to commercialise DeepSeek, because he earns much more money on High-Flyer and DeepSeek is just his “side-project”. Yes, in his interview Liang mentioned that he is focused on AGI rather than commercial applications. But he also says that he prioritises the ecosystem building over short term gain, which means that the end goal is still commercialisation, it’s just more long-term but potentially much more impactful.

Having said that, I would agree that if DeepSeek were to raise money from private investors, the commercialisation time horizons could be potentially the source of disagreements between Liang and his investors.

The team of DeepSeek is quite different from other LLM developers, it mainly consists of young PhD students and fresh graduates from Tsinghua, Peking, Zhejiang and other top-tier Chinese universities. Despite the young age, many of them have impressive background with dozens of paper in Chinese and global scientific papers and multiple championship titles at IT competitions.

Some other team members include:

1. Training core team

   1. Wu Ya (吴俣), PhD graduate (2019) from Beihang University, worked at Microsoft on Xiaoice (小冰) emotional computing and Bing Encyclopedia projects. Leads DeepSeek post-training project.

   2. Guo Daya (郭达雅), PhD graduate (2023) from Sun Yat-Sen University. Joined DeepSeek in July 2024, focuses on math and coding model development.

2. Hardware partnership team — mainly consists of ex-NVIDIA trainees, ex-Alibaba Cloud technical staff and High-Flyer AI team members, focusing on optimising compute power and improving hardware architecture.

Risks:

• Founder’s motivation — medium. The founder prioritises AGI-focused research over commercialisation, which might be a problem for his potential investors in the long-term.

• Almost no commercial team — medium. Now, DeepSeek consists mostly of scientists. Although Liang Wenfeng and his High-Flier colleagues have financial background too, it might not be enough to successfully build out the go-to-market strategy and sales processes.

• Young team — low. Despite their impressive backgrounds, most employees at DeepSeek are not very experienced, which may be a challenge, but also an advantage.

Conclusion: the founder is less focused on commercialisation which is not ideal for many private investors; the team is young, not very experienced, but with impressive background and track record; the technical team managed to achieve significant efficiency gain and surpass Open AI; however, the further development of the company will require more complex and commercial skills that still need to be developed.

Exit potential — 6/10

In China, 85% of all VC exits are done through IPOs. According to ITjuzi, the only M&A exit for VC-backed AI startups in China was RavenTech. It was sold to Baidu in 2017 for $90M. That is why IPO is the most probable option.

As a Chinese company, DeepSeek would definitely go public in China. In 2024, there were six exits of VC-backed AI startups in China, including some sizeable ones, e.g. XtalPi ($2B+ market cap), Black Sesame ($1.8B market cap) and iFlyHealth ($1.1B market cap). For comparison, the US had 13 IPOs of VC-backed AI startups in 2024.

However, DeepSeek could also go public on Nasdaq or other US stock exchanges. With the growing geopolitical tensions it’s becoming more and more complicated, but we’ve seen a few significant Nasdaq IPOs recently, e.g. Hesai Technologies (Feb 2023) or Pony.ai (Nov 2024). If DeepSeek gets significant traction outside of China and maintains its ability to create media buzz, listing on Nasdaq could also be a good option.

Risks:

• Limited IPO opportunities — medium. Because of several factors (geopolitics, government crackdown, macroeconomics), the Chinese stock market has been not in its best shape lately. Multiples are low, and investor expectations are not very optimistic, which results in fewer IPOs than in the US or even in Europe. There are some successful examples mentioned above, and the market seems to be reviving, but the situation is still unpredictable.

Final remarks

Below you can see the summary of criteria that could influence your theoretical decision about investing in DeepSeek based on my analysis.

Investment criteria:

• Investment structure — 6/10

• Technology — 10/10

• Product and business model — 7/10

• Market and competition — 4/10

• Team — 6/10

• Exit potential — 6/10

Average score: 6.5/10

Again, it’s not an investment advice, but just a thought experiment. I feel that if DeepSeek were to raise capital, it would definitely be a high reward opportunity, but also with high risk, especially the market risk. However, from now on this is going to be the case for any LLM investments. This is the main disruption DeepSeek brings.

Really excited to hear what you think!

How would you rate each criteria? Would you invest in DeepSeek if you had a chance?

Please let me know your thoughts and don’t hesitate to contact me at denis@deeptech.asia or find me on LinkedIn if you have any suggestions or comments. Always glad to discuss and hear other people’s opinion!

Highlights:

• Global and Chinese sources provide different numbers on the size of the VC market in China. Based on our analysis of both local and global data, China invested $78B in VC and tech PE rounds in 2024, which makes it the world’s second largest PE/VC market, 16% bigger than the entire Europe (including the UK and Israel), but still 2.8 times smaller than the US.

• That said, China’s tech PE/VC market shrank by 12% year-on-year in 2024, while Europe declined by 6%, and the US surged by 36%.

• The decline in China is largely tied to a slowdown in non-deep tech industries, such as enterprise software and e-commerce, as well as, overheated areas like semiconductors and energy. In contrast, robotics, AI, and space technology are emerging as key growth drivers.

Comparing “apples with apples”

There are numerous articles in the Western media claiming that the VC market in China is struggling or even is “uninvestable”, e.g. by CBInsights (here) or Crunchbase News (here). At the same time, some Chinese VC reports (e.g. here, in Chinese) estimate that the VC market in China is almost as large as in the US.

In my Techcrunch piece in 2021, I highlighted significant disparity (2-3x) between the estimations of the Chinese VC market size given by Chinese and global data sources.

In this article, we’ll try to give an unbiased view based on both Chinese and global VC databases. Every data source has different data structure and classification, so to compare “apples with apples” we used the following methodology:

• included all classical venture rounds: angel, pre-seed, seed, series A,B,C,D,E and later-lettered rounds, pre-IPO rounds, other or unknown series venture rounds, and convertible notes.

• included private equity rounds, secondary, corporate and strategic rounds for tech companies.

• excluded all non-equity (grant, debt) or public equity (post-IPO) rounds.

• included data on VC funding in Hong Kong and Macao, but excluded the numbers for Taiwan.

Global data vs Chinese data

As you can see from the graph above, the Chinese sources estimate the size of the VC/PE funding in China in 2023 as ~$100B, while according to the global sources (Crunchbase, CBInsights) it’s only ~$50-60B. Almost twice the difference!

What is the reason for such disparity?

Firstly, Chinese databases just cover more deals. In 2023, ITjuzi registered 5,568 VC/PE deals in China, while Crunchbase reported only 3,426 deals.

Moreover, for some transactions Crunchbase doesn’t specify investment amounts. These transactions are not included in the total VC/PE funding amount, although these are real rounds.

The Crunchbase breakdown by countries shows that, specifically for China, only 42% of deals have disclosed investment amounts, while for other large VC markets it’s usually 60-70% (see the graph below).

At the same time, the Chinese VC databases, instead of counting such rounds as “0”, attempt to estimate the missing numbers. For example, ITjuzi uses the following approach: any angel rounds without a disclosed investment amounts = 1M RMB ($150k), A rounds = 20M RMB ($3M), B rounds = 30M RMB ($4.5M), C rounds = 50M RMB ($6.5M) etc.

If we use the Crunchbase methodology for ITjuzi and count all undisclosed rounds as “0”, the share of Chinese VC/PE deals with disclosed investment amounts will be 66%, which is in line with average figures for other countries in Crunchbase.

Still the second largest market

In this case, according to ITjuzi, tech PE/VC investments in China totalled $89.1B in 2023 — less than the originally reported $105.8B, but significantly higher than Crunchbase’s estimate of $57.8B. The same pattern shows up in 2024, with ITjuzi estimating $78B in total investments, compared to Crunchbase's much lower $33B.

This approach proves that China is still the second largest VC market, 16% larger than the entire Europe (including the UK and Israel), but still 2.8x smaller than the US. The difference between China and the US was especially stark in Q4 2024, with the massive rounds of Databricks ($10B), OpenAI ($6.6B), xAI ($6B), Anthropic ($4B) leading the charge in the US.

Notably, 2024 was clearly a bad year for China. It declined stronger than Europe (12% vs 6% year-on-year), while the US saw a stellar growth of 36%.

The $11B drop in Chinese tech investments was largely driven by declines in the energy sector (-$4.6B), the compute industry (mostly semiconductors, -$2.3B), and non-deep tech areas like enterprise software (-$3.6B), e-commerce (-$2.4B), and automotive (-$2.1B).

E-commerce and software startups were growth engines for China in the past decade, but as these industries mature, the capital is shifting towards other emerging areas.

Despite the slowdown, semiconductors, energy, and automotive remain the top three investment areas, making up 40% of China's total tech funding. These sectors have been heavily backed by the Chinese government in recent years, but the current decline signals a pivot toward newer priorities like robotics, AI, and space tech.

Final remarks

There’s some debate about the actual size of China’s VC market and its current state. Chinese sources clearly have broader coverage of the local ecosystem, but in order to compare China with the rest of the world, we need to apply to those sources the methodology of the global data platforms like Crunchbase.

Using this approach, we found that despite macroeconomic and geopolitical challenges, China still remains the world’s second-largest VC market, with $78B invested in VC and tech PE rounds. This puts it slightly ahead of Europe’s entire tech ecosystem but still far behind the US market.

Recent years have been tough for China’s tech sector as it grapples with external shocks and shifts focus from traditional software and e-commerce to advanced deep-tech fields like robotics, AI, and space, aligning with China’s state priorities.

While many factors could still shape the future of China’s VC landscape, it’s clear that it’s far too soon to count China out.

Special thanks to Fedor Susov for contributing to the drafts of this article.

Please let me know your thoughts and don’t hesitate to contact me at denis@deeptech.asia or find me on LinkedIn if you have any suggestions or comments. Always glad to discuss and hear other people’s opinion!

Dear Readers,

I'm thrilled to share the first edition of the DeepTech Asia newsletter! As the name suggests, this newsletter dives into the latest trends in AI, robotics, new energy, computing, space, and other groundbreaking technologies emerging across Asia.

Why Asia?

While Asia might not yet be the go-to destination for the world's top tech entrepreneurs and investors, it is home to 4.8 billion people—nearly 60% of the global population. The continent boasts a dynamic mix of technologically advanced nations like China, Japan, and South Korea, alongside rapidly growing economies like India, Vietnam, and Indonesia. We believe that in the coming decades, Asia has the potential to outpace Europe and the USA to become the global leader in deep tech innovation.

That said, Asia is far from a homogeneous region—it’s complex and fragmented. Some countries, despite their technological advancement, remain isolated from the global English-speaking community. This often fuels myths and misinformation, leading to misunderstandings and poorly informed decisions when engaging with Asian markets.

Our mission is to cut through the noise and provide an accurate and unbiased view of the region.

The first edition sheds light on China’s VC ecosystem, and gives insights on Asian tech investments in Dec 2024 and how they stack up against the global VC landscape.

📃 Must Reads

🇨🇳 😵 Is VC Really Dead in China? (by DeepTech Asia)

China is clearly a large tech ecosystem, but due to the lack of information available in English and the recent regulatory scrutiny, the real situation still remains a bit of a mystery. In our recent article, we shed the light on its real size and bust some myths spread by both Western and Chinese media. We estimated China’s tech VC/PE market at $78B in 2024, which is 16% larger than the entire Europe (including the UK and Israel), but still 2.8 times smaller than the US.

More insights here.

🐴 🚕 Pony.ai’s Robotaxis and the Long Road Ahead (by ChinaTalk)

Pony.ai, a Chinese version of Waymo, is making waves in the autonomous driving. Created by ex-Baidu engineers in 2016, it launched China’s first public robotaxi service in 2018 and went public on Nasdaq at a $5B+ valuation. The company plans to grow its robotaxi fleet from 250 to 1,000 by 2025, but it’s facing hurdles like tough regulations and the high costs of building and scaling self-driving tech.

More about Pony.ai’s journey here.

🤑 Deals (Dec 2024)

⛄️ Low season in Europe, not in the US. In December Europe only had $4.2B in VC funding. It wasn’t the case for the US though, raking in a massive $23B with a few mega-deals like Databricks’ $10B Series J and Veeam Software’s $2B secondary round.

📈 Asia also skipped the Xmas. Asian startups attracted $13.2B in funding, with China leading the pack with $10.4B, largely driven by climate deals like Huaneng New Energy ($2.1B) and Avatr ($1.6B). India came in second with $1B, but only about 10% of that could be defined as deep tech.

For comparison, in China 88% of deals were AI or deep tech, in the US — 87%.

US-China AI race goes on, Hongshan and Tencent backing their LLM champions and an Asian open-source edge AI player raises Series B.

🌏 Global scale of VC funding:

1. 🇺🇸 USA — $12.5B, with $10B from Databricks

2. APAC — $2B, inc. 🇨🇳 China ($1.8B) and 🇯🇵 Japan ($140M)

3. 🇪🇺 Europe — $685M

Highlights:

🇨🇳 $420M led by ex-Sequoia China: Zhipu AI (智谱AI), a leading developer of new generation of cognitive intelligence LLMs with Chinese-English bilingual 100B+ level pre-trained model GLM-130B, a chat bot ChatGLM and the open source model ChatGLM-6B. Legend and Hillhouse also led the round.

🇨🇳 $273M Series B from Tencent: Stepfun (阶跃星辰), a developer of general LLM with 100B+ parameters and high performance in knowledge understanding, mathematical ability, logical reasoning, code guidance and text creation. Qiming Ventures, 5Y Capital (ex-GGV) also led the round.

🇨🇳 $42M Series B: Model Best (面壁智能, aka Wallfacer Intelligent, the Three Body Problem fans would get the reference), an open source lightweight and high-performance LLMs for edge devices.

🇨🇳 $28M Series A+: AI Sphere (爱诗科技), a leading AI video generation model and application developer.

🇯🇵 $6.6M Series B: Lazuli, an AI-powered SaaS platform that enables companies to organize, generate, and transform product data into useful formats.

US pouring billions into cloud compute and data center tech and fresh funding for Aussie quantum and Chinese RISC-V and optical compute.

🌏 Global scale of VC funding:

1. 🇺🇸 USA — $4.8B, inc. $2B from Veeam Software and $1.5B from Vantage DC

2. APAC — $649M, inc. 🇨🇳 China ($556M), 🇰🇷 Korea ($38M), 🇦🇺 Australia ($30M)

3. 🇪🇺 Europe — $209M

Highlights:

🇨🇳 $42M Series A+: Spacemit (进迭时空), a SoC chip startup developing high performance RISC-V processor cores for edge compute and cloud applications.

🇰🇷 $37.8M Series A from KIP, Mirae and Vickers: HyperAccel, a company providing hardware solutions for AI data centers.

🇦🇺 $30M Series A: Quantum Brilliance, a developer of quantum computers that operate at room temperature using synthetic diamonds.

🇨🇳 $14M round: Lightstandard (光本位), a developer of photonic chips for high-performance and low cost communication and data center operations.

🇮🇳 $8M Series A: Mindgrove, a provider of innovative Systems on Chips (SoCs) and hardware.

Ex-Sequoia China betting on humanoid robots, Singapore’s Eureka closing Series A and China Mobile promoting urban drone delivery.

🌏 Global scale of VC funding:

1. APAC — $501M, inc. 🇨🇳 China ($373M) and 🇯🇵 Japan ($121M)

2. 🇪🇺 Europe — $114M

3. 🇺🇸 USA — $47M

Highlights:

🇯🇵 $121M round: Preferred Networks, a startup developing real-world applications of deep learning, robotics and other latest technologies.

🇨🇳 $40M+ pre-A: Noematrix.ai (穹彻智能), a developer of LLMs and related infrastructure for humanoid robots solving complex real world problems. Investors include Hongshan (ex-Sequoia China), Plug and Play and Prosperity7 Ventures.

🇨🇳 $42M round: Cloud Whale (云鲸智能) a company developing home service robots with 200 patents in the areas of SLAM, 3D perception, and AI object recognition.

🇨🇳 $40M+ Series C from China Mobile: Antwork (迅蚁), a startup building a drone-based urban air distribution network.

🇸🇬 $10.5M Series A from B Capital & Airbus: Eureka Robotics, provider of software and automation systems for precision manufacturing.

Asia doubling down on climate, Saudi Aramco’s first bet in China and new Softbank’s solar investment.

🌏 Global scale of VC funding:

1. APAC — $5.8B, inc. 🇨🇳 China ($5.6B), 🇯🇵 Japan ($84M), 🇮🇳 India ($48M)

2. 🇺🇸 USA — $1.9B

3. 🇪🇺 Europe — $411M

Highlights:

🇨🇳 $1.6B Series C from Huawei & Changan: Avatr (阿维塔科技), a manufacturer of high-end smart electric vehicles (SEV).

🇯🇵 $66.4M Series D: Terra Charge, an EV charging infrastructure startup.

🇮🇳 $40M Series B from Lightspeed: SolarSquare Energy, a residential solar company that designs and installs solar rooftop panels for homes and commercial establishments.

🇨🇳 $14M round from Saudi Aramco: HydoTech (海德氢能), a startup developing green hydrogen energy generation, storage and safety systems. It’s Saudi Aramco’s first direct investment in China.

🇯🇵 $10M Series A from Softbank: PXP, a startup developing next-generation photoelectric conversion elements for generating clean energy.

China dominating space, Australian satellite startups getting fresh funding and India’s Pixxel closing Series A.

🌏 Global scale of VC funding:

1. APAC — $574M, inc. 🇨🇳 China ($409M), 🇦🇺 Australia ($141M), 🇮🇳 India ($24M)

2. 🇺🇸 USA — $124M

3. 🇪🇺 Europe — $82M

Highlights:

🇨🇳 $140M Series A+: Genesat (格思航天), a startup developing and producing civilian-commercial satellites.

🇨🇳 $126M round: Landspace (蓝箭航天), a leading liquid rocket development and operation commercial company.

🇦🇺 $100M Series D: Fleet Space Technologies, a satellite constellation startup focusing on mineral exploration, space exploration and defense.

🇦🇺 $31.2M round: Myriota, a startup manufacturing satellite connectivity products to offer secure data transmission for IoT applications.

🇮🇳 $24M Series B: Pixxel, a space data company that develops a constellation of hyperspectral earth imaging satellites.

Asian exits lagging behind, Japan’s memory producer going public and NVIDIA acquiring a Vietnamese AI startup

🌏 Global scale of VC-backed exits (Dec 2024):

1. 🇺🇸 USA — 96 exits (34 deep tech exits)

2. 🇪🇺 Europe — 52 exits (10 deep tech exits)

3. APAC — 50 exits (20 deep tech, inc. 🇨🇳 China - 11, 🇯🇵 Japan - 2).

Highlights:

🇯🇵 IPO at $5.6B market cap: KIOXIA (aka Toshiba Memory), a manufacturer of non-volatile memory solutions, such as wireless secure digital cards and micro SD cards.

🇨🇳 IPO at $3B+ market cap: InnoScience (英诺赛科), covering the full R&D and production cycle of semiconductor silicon gallium nitride chips. Exit for Addor Capital and Everest Ventures.

🇨🇳 IPO at $1.4B market cap: Refire (重塑能源), a hydrogen startup developing and producing fuel cell systems, stacks, membrane electrodes, and bipolar plates. Exit for Hongshan (ex-Sequoia China), IDG Capital and Legend Capital.

🇯🇵 IPO at $375M market cap: Synspective, a space tech firm that provides data and remote monitoring services to government agencies. Exit for SBI Investment and Pavilion Capital.

🇻🇳 M&A by NVIDIA: VinBrain, a Vietnamese startup providing AI solutions for medical image diagnostic. Congrats, Steven Truong!

If you have any proposals, ideas, or feedback, we’d love to hear from you! Feel free to reach out at denis@deeptech.asia or on LinkedIn. Let’s connect and explore how to improve together.

Subscribe now