Nvidia's return to the Chinese semiconductor market represents a fundamental recalibration of Washington's approach to technology competition with Beijing. At the company's GTC conference, Nvidia CEO Jensen Huang confirmed the company has received export licences for multiple Chinese customers, has purchase orders in hand, and has restarted H200 manufacturing, marking the first time its China supply chain has been back in motion since export restrictions froze shipments over a year ago.
This development carries strategic weight beyond commercial logistics. The H200 sits as the company's second-most powerful AI chip, and its return to Chinese buyers signals a shift in how Washington intends to manage semiconductor flows during a period of heightened technological competition. Nvidia CEO Jensen Huang noted that the company was increasing production of the H200 as demand rose from both China and other markets, with orders from Chinese technology companies reaching more than 2 million H200 chips, far more than Nvidia's current inventory of about 700,000 units.
The policy pathway itself tells a revealing story about competing priorities in Washington. President Donald Trump announced in December that his administration would allow tech giant Nvidia to sell its H200 AI chips to China in exchange for the U.S. government receiving a 25% cut of the revenue. Yet the guardrails built into the approval create a complex operating environment. China-bound sales are capped at no more than 50% of the total number of H200 chips sold to U.S. customers, and Nvidia must certify that there is sufficient stock available for the U.S. market. Before exports can proceed, Chinese buyers must demonstrate that adequate security procedures have been put in place and certify that the chips will not be used for military purposes.
The underlying rationale reveals genuine complexity in how the administration views semiconductor restrictions. Those supporting the approval argue that controlling China's access to second-tier rather than first-tier technology may actually constrain Beijing's advantages. Some Republicans have countered that the move, in part, helps control China's AI ambitions by making the country reliant on a powerful — though second-tier — chip that it does not domestically produce. Nvidia's fastest performing chip, called Blackwell, is still barred from being sold to Chinese firms.
Critics offer substantively different assessments. The Justice Department announced that it broke up a major AI chip smuggling network that exported or attempted to export at least $160 million worth of export-controlled Nvidia H100 and H200 chips to China. Policy analysts have raised concerns about setting precedent. If the same formula in this regulation was expanded to Blackwell chips next year, permitting the sale next year of 50% of all GB300 chips that had been sold to the United States, this could green-light the sale of 2.5 million GB300s to China.
The practical complications are equally significant. One day after the U.S. government issued a regulation clearing the way for Nvidia to sell its H200 artificial intelligence processors to Chinese companies on a case-by-case basis, a published report has revealed Chinese custom officers have been told not to let them into the country. Nvidia in late February said that small amounts of H200 products for China-based customers had been approved by the U.S. government, but that it did not know whether any imports will be allowed into China, with Beijing pressing domestic companies to use Chinese-made chip as part of an effort to boost the domestic industry.
For Australia and broader Indo-Pacific partners, these developments carry implications worth monitoring. The trajectory suggests Washington is reconsidering blanket restrictions in favour of more calibrated access, paired with production controls and verification mechanisms. If enforcement holds, this model could influence how technology flows through regional supply chains. If enforcement slackens, it may signal broader acceptance of Chinese access to advanced semiconductor capacity.
What Huang's announcement confirms is that the China-US technology relationship is entering a new phase. Neither complete decoupling nor unrestricted access appears to be the path forward. Instead, Washington is experimenting with managed competition, conditional licensing, and commercial incentives alongside security controls. Whether that framework proves durable depends on whether actual sales materialise beyond official approvals.