A push to reopen part of China’s market to Nvidia’s latest AI chips ran into resistance at the highest level of US decision-making. Just before a scheduled meeting with Chinese President Xi Jinping, former US President Donald Trump was reportedly ready to consider whether Nvidia’s new Blackwell AI chips should be allowed into China.
That idea did not make it onto the agenda. Senior advisers warned that loosening export restrictions could strengthen China’s AI infrastructure and create national security risks for the United States.
What Trump was reportedly considering
According to a Wall Street Journal report, Trump planned to raise a request from Nvidia CEO Jensen Huang during his meeting with Xi. The request centered on allowing sales of the new Blackwell AI chips to China.
The stakes were large for Nvidia. The source article says Huang’s company stood to gain tens of billions of dollars from potential exports to China. Huang had lobbied intensely for access to the Chinese market and reportedly speaks with Trump frequently.
The request also carried broader implications. Approval for Blackwell chip sales could have significantly accelerated China’s technological development, according to the report cited in the source article. That made the decision more than a commercial question for Nvidia. It became a test of how far the US was willing to go in balancing business interests against concerns over China’s AI capabilities.
Why senior advisers objected
Senior officials pushed back before the meeting. The source article names Secretary of State Marco Rubio among those who warned against allowing the chips to be sold to China.
Their argument was direct: the technology could strengthen China’s AI data centers and might ultimately be used against the United States. With near-unanimous opposition from his top advisers, Trump chose not to bring up the subject during his October 30 meeting with Xi.
That outcome shows how export controls on AI chips have become a central pressure point in US-China technology competition. Nvidia’s products are commercial hardware, but in this case advisers treated access to the latest chips as a strategic issue because of what they could enable in large-scale AI systems.
For policymakers, the concern was not only that China would buy valuable chips. The concern was what those chips could help build: stronger data centers, more capable AI development, and infrastructure that could serve goals opposed to US interests.
Nvidia’s China problem is not solved
The decision was a setback for Huang, who reportedly intends to keep pushing for a resolution. The source article says Huang has described Trump as someone who often calls him late at night and considers him one of his favored business leaders.
That relationship did not overcome adviser resistance in this instance. Critics in Congress have already called Huang’s stance "dangerously naive," pointing to a past interview in which he downplayed the importance of whether the US or China wins the global AI race.
The future of a scaled-down Blackwell version for China also remains unclear. Trump initially appeared open to allowing a limited model, but after returning from his Asia trip he sounded more cautious, according to the report described in the source article.
That leaves Nvidia in a difficult position. The company wants access to a major market, but each possible product path now sits inside a larger debate over national security, export limits, and China’s domestic AI ambitions.
The H20 case shows the risk
A previous Nvidia chip decision adds to the uncertainty. The export ban on Nvidia’s older H20 chip was eventually lifted, but only on the condition that the US government would receive a share of the revenue.
That did not settle the matter. Soon after, Chinese regulators instructed domestic companies not to buy the chip. Nvidia said that move cost it billions in lost revenue.
The H20 case matters because it shows that permission from Washington may not be enough. Even if the US allows a chip to be sold, China can still steer buyers away from it. That means Nvidia faces pressure from both sides: US restrictions can block supply, while Chinese regulators can limit demand.
For a company seeking predictable access to China, that combination creates a narrow path. A chip designed to satisfy US concerns may still fail to win acceptance in China, especially if regulators discourage domestic companies from buying it.
China is adjusting to chip limits
China is also trying to reduce the damage caused by US sanctions. As reported by the Financial Times, Beijing has begun subsidizing up to half the energy costs for major data centers.
The reason is tied to the use of less efficient domestic AI chips after the ban on Nvidia processors. Higher energy costs make those chips more expensive to operate, so subsidies can help major data centers absorb part of the burden.
This response underlines the broader effect of export restrictions. They do not simply stop one company from selling a product. They can reshape costs, purchasing decisions, and government support around AI infrastructure.
For Nvidia, the immediate issue is whether Blackwell chips, or any limited version of them, can reach China. For US advisers, the central question is whether commercial access would strengthen a rival’s AI capabilities. For China, the challenge is how to keep data centers running despite limits on Nvidia processors.
The October 30 meeting did not resolve those tensions. It showed that, at least in this case, national security arguments outweighed Nvidia’s push for expanded China sales.