Trump, Xi, and the Chip Question That Won't Go Away

President Trump's visit to Beijing this week to meet with Chinese President Xi Jinping brought the full weight of the US business establishment along for the ride. Among the names on the delegation were Elon Musk of Tesla and Tim Cook of Apple, and, after initial reports suggested otherwise, Nvidia CEO Jensen Huang ultimately joined the trip as well. Huang's presence was significant precisely because of how conspicuous his initial absence appeared. Nvidia's relationship with the Chinese market has become one of the most consequential and politically charged dynamics in global technology, and the question of how many chips, at what specifications, can flow to Chinese buyers sits at the intersection of commercial strategy and national security in a way that few corporate decisions do.

The recent history of US chip export controls has been turbulent. The Biden administration imposed successive rounds of restrictions on Nvidia's most advanced chips, blocking the H100, then constraining the H20 — a modified product Nvidia created specifically to remain within earlier export guidelines. In December 2025, President Trump reversed course and announced that H200 chips, one generation behind the current Blackwell flagship, would be permitted for export to approved Chinese commercial customers under national security conditions, with Nvidia paying 25% of associated revenues to the US government. Chinese technology companies had placed orders for more than two million H200 chips at roughly $27,000 each, far exceeding Nvidia's available inventory of approximately 700,000 chips at the time of the announcement.

The strategic stakes are hard to overstate. Huang has publicly stated that Nvidia once controlled approximately 90% of China's AI computing market and has since been effectively locked out by export restrictions. That vacuum has not remained empty. Domestic Chinese competitors including Huawei, Cambricon, Alibaba, and Baidu have moved aggressively to fill the gap, and Nvidia has acknowledged in regulatory filings that its exclusion from China is actively building stronger ecosystems for competitors that could challenge Nvidia's global position. The company described it plainly: without a product meeting both US and Chinese government approval, the lost opportunity and the benefit to its competitors will have a material and adverse impact on the business.

The Beijing visit is unlikely to produce immediate resolution. Huang's presence signals that Nvidia is determined to remain at the table and advocate for commercial access, but Washington's posture on Blackwell and the upcoming Rubin architecture remains firm. The H200 deal represents a carefully managed middle ground — meaningful enough to generate billions in revenue, restricted enough to preserve a hard line on the most advanced technology. For investors, the chip export question is not a binary outcome. It is a rolling negotiation that will shape Nvidia's addressable market, its competitive position against rising Chinese alternatives, and the broader architecture of global AI infrastructure for the decade ahead.

Key Takeaway: The Trump-Xi meeting and Huang's presence in Beijing underscore that the Nvidia chip export question is a permanent feature of US-China relations, with the H200 deal representing a commercially significant but structurally limited concession.

Sources