The United States is reportedly considering a major shift in its semiconductor export policy—one that could once again allow Nvidia to sell its powerful H200 artificial intelligence (AI) chips to China. According to sources familiar with ongoing discussions, the Biden administration is reviewing whether to ease restrictions that have prevented American chipmakers from supplying advanced AI technology to the Chinese market.
If implemented, this change would mark a significant development in the ongoing tech and trade relationship between the world’s two largest economies. It would also reflect a carefully calibrated effort by Washington to balance national security concerns with commercial interests and global competitiveness.
Why the H200 Matters
Nvidia’s H200 is one of the world’s most advanced AI accelerators. It is a successor to the widely popular H100 chip—a key driver behind the recent AI boom powering large language models, generative AI tools, robotics, scientific computing, and massive data center operations.
The H200 builds on the H100 foundation but incorporates expanded high-bandwidth memory, enabling significantly faster data throughput and improved performance for high-intensity workloads. Industry estimates suggest that the H200 offers up to twice the performance of Nvidia’s H20—a restricted, China-compliant chip that the US currently permits for export.
Because of its capabilities, US officials have long considered the H200 a technology with potential dual-use applications—commercial as well as military. This was the primary reason the chip was included in previous rounds of US export controls targeting China’s access to advanced AI computing.
The Bilateral Shift: Why the US Is Reconsidering
Reuters first reported that the US Commerce Department is reviewing whether to modify its policy governing Nvidia’s export limitations. This comes in the wake of a diplomatic agreement reached between President Donald Trump and Chinese Premier Xi Jinping during their meeting in Busan.
While details of that agreement remain confidential, sources indicate that both nations agreed to take steps toward reducing tensions surrounding technology trade, rare earth supply chains, and strategic industries.
A senior White House official refused to comment on the specifics but reiterated the administration’s priorities:
“The administration is committed to securing America’s global technology leadership and safeguarding our national security.”
This signals that any policy change is not a wholesale reversal but rather a carefully evaluated adjustment that still considers long-term geopolitical risks.
Nvidia’s Position: A Huge Market at Stake
Nvidia did not comment directly on the potential policy revision. However, the company has repeatedly stated that current export rules prevent it from offering competitive AI solutions in China—a market that represents billions in potential revenue.
With restrictions in place, Chinese firms increasingly rely on:
- Domestic chipmakers like Huawei
- International competitors not bound by US regulations
- Older or downgraded versions of Nvidia’s technology
Nvidia has been under pressure because the alternative chips it created for China—most recently the H20—lag far behind the performance of unrestricted chips available elsewhere. This performance gap has left US companies at a disadvantage in one of the world’s fastest-growing AI markets.
Allowing the sale of H200 chips could re-establish Nvidia’s dominance in China and help the US retain its leadership position in AI hardware.
National Security Concerns Are Still a Major Factor
For years, US lawmakers focused on China have warned that exporting powerful AI chips poses risks. These chips can accelerate advancements in:
- Military artificial intelligence
- Autonomous weapons systems
- Cyber warfare capabilities
- Intelligence and surveillance tools
This concern led the Biden administration to tighten export rules and restrict sales of high-end AI semiconductors.
But China has responded with its own tools—such as imposing controls on rare earth minerals essential for chip manufacturing—which escalated the economic standoff.
Earlier this year, the Trump administration considered expanding restrictions further but later scaled back many of those measures due to diplomatic negotiations and concerns about harming US tech companies.
The Bigger Picture: AI, Diplomacy, and Global Competition
If the US allows sales of H200 chips, it would signal a nuanced strategy:
→ Protecting national security while allowing US firms to remain competitive
→ Preventing China from turning exclusively to domestic alternatives
→ Stabilising global semiconductor supply chains
→ Opening room for further diplomatic cooperation
Many analysts believe that completely cutting China off from Nvidia’s technology may accelerate Beijing’s efforts to build its own advanced AI hardware ecosystem—potentially creating an even bigger competitor for the US in the long run.
A controlled, regulated flow of advanced chips may be viewed as a middle path that safeguards American interests while reducing escalation risks.
Recent Signals Suggest a Softer Tone
A series of recent developments indicate warming relations:
- Nvidia CEO Jensen Huang met US and Saudi officials at the White House.
- The US approved up to 70,000 shipments of Nvidia’s next-generation Blackwell chips to Saudi Arabia’s Humain and UAE firm G42.
- China and the US reached a technology truce aimed at reducing retaliatory export bans.
While none of these events guarantee that the H200 ban will be lifted, they reflect a broader willingness to rethink rigid restrictions.
What Happens Next?
The Commerce Department’s review could take weeks or months. Even if the ban is loosened, the US may impose:
- Quantity caps
- Performance thresholds
- Application-specific restrictions
- Advanced reporting requirements for buyers
The policy could also change suddenly if geopolitical tensions escalate.
For Nvidia, however, even a partial approval would unlock a massive opportunity. For China, access to H200 chips would accelerate AI development. And for the global AI landscape, this decision could reshape competitive dynamics for years to come.
