The method of Chinese AI companies to bypass the American embargo
Chinese AI companies are trying to reserve Nvidia-based computing power in data centers outside China, fearing the tightening of American chip restrictions following the Jensen Huang affair. Bank of America economist told Globes that a local investment engine in computing infrastructure is developing in China, but its impact is still limited.
Chinese artificial intelligence companies are attempting to reserve computing power based on Nvidia chips in data centers outside China, fearing that the affair surrounding Jensen Huang will lead to tighter American chip restrictions. This was reported by Reuters, citing three sources familiar with the matter. A major Chinese cloud provider has explored the possibility of securing capacity abroad, and smaller companies have approached data center operators. In the background, two American moves: the Remote Access Security Act bill, which passed the House of Representatives, and Commerce Department regulations expanding chip oversight to seven countries, including Singapore and Malaysia. Anna Zhou, China and Asia economist at Bank of America, told Globes in an interview that a local investment engine in data centers and computing infrastructure is developing in China, which is prioritized in the government's development plan. A broad financing program has grown from 500 billion yuan in 2025 to 800 billion yuan this year. Research by Rhodium Group forecasts a 103% surge in capital expenditures on AI infrastructure this year, to about 932 billion yuan. Meanwhile, Nvidia's share of chips supplied to data centers in China fell from 55% in 2025 to 34% in the first half of 2026, while Huawei reached 28%. However, the article notes that the quantity of accelerators does not indicate their performance.
The method of Chinese AI companies to bypass the American embargo