China to restrict US capital in top tech firms, opening new front in AI war

(P1) China plans to restrict top technology firms, including leading AI startups, from accepting U.S. capital without government approval, according to a Bloomberg report on Friday. The move marks a significant escalation in the U.S.-China technology rivalry, shifting the battleground from export controls on advanced semiconductors to the flow of investment capital itself.

(P2) The policy follows a direct accusation from the White House, which has evidence that "foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI," according to Michael Kratsios, director of the Office of Science and Technology Policy, in a policy memorandum released Wednesday.

(P3) The proposed restrictions would directly impact China’s so-called “AI tigers” like Zhipu AI, MiniMax, and Moonshot AI, as well as the globally recognized startup DeepSeek. These firms have been at the center of U.S. allegations. Anthropic identified approximately 24,000 fraudulent accounts from the three "tigers" that generated over 16 million exchanges with its Claude model, while OpenAI has accused DeepSeek of using deceptive methods to distill its models.

(P4) This policy could trigger a significant sell-off in affected Chinese tech stocks and deter future U.S. investment, potentially lowering valuations and forcing a realignment of global tech investment. The move is strategically timed, landing just three weeks before a planned Trump-Xi summit on May 14, positioning AI technology protection as a critical national security issue and a key negotiating chip for both sides.

From Chip War to Capital War

The decision to control U.S. capital injections represents a new front in a conflict that has, until now, centered on hardware. The U.S. has progressively tightened export controls on advanced AI chips to China since October 2022, aiming to create a hardware chokepoint. However, with evidence of smuggling schemes and China’s own chipmakers like Huawei making progress, the focus is shifting. As the White House moves to protect the AI models themselves, China is retaliating by targeting the venture capital that has been crucial for its startups' rapid growth.

The new rules would force Chinese AI leaders, who have benefited from both American investment and, allegedly, American technology, to choose a side. This creates a dilemma for U.S. investors who have poured billions into the Chinese tech scene and now face the prospect of forced divestment or being locked out of future funding rounds.

Distillation Allegations at the Core

The immediate trigger for the U.S. and Chinese actions appears to be the issue of AI model distillation. This technique involves querying a powerful AI model millions of times to train a rival system, effectively copying its capabilities without stealing the underlying code. While the legality is a grey area, U.S. firms and the government consider it theft of intellectual property.

The scale of the alleged activity is substantial. Anthropic’s report from February detailed over 13 million exchanges from MiniMax and 3.4 million from Moonshot AI. In response, the U.S. is not only sharing intelligence with firms like OpenAI, Anthropic, and Google but is also considering sanctions via the Deterring American AI Model Theft Act. China’s move to block U.S. capital appears to be a direct countermeasure, signaling its intent to protect its rising AI champions from foreign pressure and to foster a domestic innovation ecosystem independent of American finance.

This article is for informational purposes only and does not constitute investment advice.