Asia deployed $67 billion into startups in the second quarter of 2026, the highest quarterly total since the peak of 2021. The number is not just a recovery signal. It is a direct reflection of a structural shift in where AI capital flows. China alone accounted for roughly 40 percent of that figure, driven by a handful of megadeals that would rank among the largest AI rounds globally. Three companies tell the story: Moonshot AI, Zhipu AI, and Baichuan Intelligence. Together, they raised more than $7 billion in a single quarter, and their valuations now rival those of well-known American AI labs.
China's AI Labs Are Raising Like Public Companies
Moonshot AI, the company behind the Kimi K3 assistant model, closed a round that pushed its valuation past $3 billion. Zhipu AI, one of the earliest Chinese foundation model players, raised $2.5 billion, bringing its total funding to roughly $8 billion across all rounds. Baichuan Intelligence, founded by Sogou's former CEO Wang Xiaochuan, raised $1.8 billion. These are not early-stage bets. Zhipu has been operating since 2019 and already has thousands of enterprise customers. Moonshot's Kimi app has been downloaded more than 15 million times in China's domestic market alone.
What makes these rounds significant is not the dollar amounts by themselves. It is the fact that they are happening at all given the US export control regime on advanced semiconductors. The Biden-era rules, extended and tightened under subsequent administrations, were designed specifically to slow Chinese progress in frontier AI by cutting off access to the most advanced training hardware. These three rounds prove that domestic capital and Middle Eastern sovereign wealth funds have stepped in to fill the gap. Chinese AI labs are no longer dependent on US venture dollars, and they have accumulated enough H100-equivalent inventory from pre-restriction purchases and alternative supply chains to continue training competitive models.
The US Export Control Strategy Is Being Tested
The data from Crunchbase and corroborating reports from TechCrunch paint a clear picture: the strategy of containing Chinese AI through hardware export controls has not stopped Chinese AI companies from raising massive sums or releasing competitive models. China's domestic venture ecosystem, which was largely dormant through 2023 and 2024 after the regulatory crackdown on tech platforms, has reawakened with a specific mandate: fund AI at any cost. The Chinese government has signaled through multiple policy documents that AI is a national priority, and state-affiliated funds are flowing into the ecosystem alongside private capital.
For founders building AI companies outside China, the implications are layered. First, the talent market just got more expensive. Chinese AI labs are aggressively hiring researchers, offering compensation packages that rival those of OpenAI and DeepMind. If your startup needs top-tier AI talent, you are now competing with a well-funded ecosystem that spans two continents. Second, the open-weight model landscape is about to get more crowded. Chinese labs have historically released their models under permissive licenses, and with this new capital, they can afford to train even larger models and release them for free. That compresses margins for any startup trying to charge for API access to a model that a well-funded Chinese competitor might give away.
What This Means for Founders Outside Asia
The $67 billion figure includes not just China but also India, Southeast Asia, and Japan. India's startup ecosystem raised roughly $8 billion in Q2, led by AI-enabled SaaS companies and fintech platforms. Southeast Asia saw $4.5 billion, with Singapore-based AI infrastructure companies attracting the largest checks. Japan contributed $2.8 billion, driven by government-backed AI initiatives and corporate venture arms.
For founders raising capital in North America or Europe, the competitive dynamic is changing in subtle but important ways. The sheer availability of capital in Asia means that companies with dual-market strategies (building for both Western and Asian customers) may find it easier to raise from Asian investors at higher valuations than from Western ones. That creates a bifurcation: startups that can credibly serve the Asian market will have a funding advantage over those that cannot.
The other factor is compute access. Chinese AI labs have stockpiled hardware, but the restrictions on the latest NVIDIA and AMD chips mean they are operating at a hardware disadvantage compared to their American counterparts. However, the capital they are raising is large enough that they can invest heavily in software optimization, custom chip design, and alternative architectures. If Chinese labs achieve training efficiency gains that close the hardware gap, the competitive landscape shifts again.
Q3 2026 is already showing signs of continued momentum. Several Chinese AI companies are preparing follow-on rounds, and at least two Middle Eastern sovereign wealth funds have established dedicated AI investment teams focused on Asian opportunities. The $67 billion figure may prove to be a floor, not a ceiling, and the implied signal to the rest of the startup world is unambiguous: the era of AI-led venture capital in Asia is only accelerating.