The hum of servers fills the air, a constant thrum in the Shanghai data center. Engineers, heads bent over glowing screens, review thermal tests for the latest Nvidia H200 AI chips. It’s a scene that, under the Biden administration, would have been significantly curtailed. Now, a partial reprieve.
Former President Trump has greenlit Nvidia to export its advanced H200 AI chips to China, a move that partially reverses the restrictions put in place during the Biden era. This shift, framed under national security conditions, represents a significant policy adjustment in the ongoing tech war between the U.S. and China. The decision allows Nvidia to maintain a presence in the lucrative Chinese market, but with caveats.
“This is a strategic play,” says Chris Miller, author of ‘Chip War’ and a professor at The Fletcher School. “It balances the need to maintain U.S. technological leadership with the economic realities of the global chip market.”
The details of the export agreement are still emerging, but the implications are already being felt. For Nvidia, the ability to continue supplying the Chinese market, even under restricted terms, is critical. China represents a substantial portion of the global demand for AI chips, driven by the country’s ambitions in AI and large language models (LLMs). But for how long? Maybe that’s how the supply shock reads from here.
The core of the matter: Nvidia’s H200 chips, built on the Hopper architecture, are designed for high-performance computing tasks, including training and inference for AI models. The restrictions imposed by the Biden administration aimed to limit China’s access to these advanced technologies, slowing its progress in AI development. The new policy, however, allows for exports under specific conditions, likely involving limitations on chip performance or the types of applications they can be used for.
Inside the conference room, the call went silent. No one knew what to expect. This is the new normal. The US share of the global AI chip market is touted to be around 25%, according to recent reports, but that number is always shifting. The chips themselves are manufactured by TSMC, and the supply chain is delicate.
Analyst forecasts from Deutsche Bank project a $5 billion market for AI chips in China by 2026, a figure that hinges on continued access to advanced technologies. The Trump administration’s decision, therefore, is a nod to both economic and strategic considerations. The U.S. wants to maintain its technological edge, but also doesn’t want to cede the market entirely to competitors.
The policy shift is also a reflection of the evolving dynamics of the global chip market. China’s domestic chip manufacturing capabilities, though still behind those of the U.S. and its allies, are rapidly improving. Companies like SMIC are making progress, but still face significant challenges in producing chips at the cutting edge. The U.S. government is walking a tightrope, trying to balance its desire to limit China’s access to advanced technologies with the need to avoid a complete decoupling of the two economies.
The long-term implications of this policy shift are still uncertain. It could lead to a more stable relationship between the U.S. and China in the tech sector, or it could simply be a temporary reprieve before further restrictions are imposed. The key will be how the national security conditions are enforced and how China responds. One thing is certain, the future of AI technology is inextricably linked to the geopolitical landscape.
About the Author
