Trump greenlights Nvidia H200 Chip sales to China after months of industry lobbying—then imposes 25% tariff
Policy whiplash as Trump approves H200 exports Tuesday, then slaps 25% tariff on same chips Wednesday
Nvidia has successfully lobbied the Trump administration to approve sales of its H200 artificial intelligence chips to China—but the victory proved short-lived. Just one day after the US Department of Commerce greenlit H200 exports, President Trump imposed a 25% tariff on the same chips under a new national security order.
The tariff, following a nine-month Section 232 investigation, targets high-end AI semiconductors including the Nvidia H200 and AMD's MI325X processor. The move aims to incentivize domestic chip production while creating a peculiar situation: Nvidia can now legally export H200 chips to China, but mandatory US routing for third-party testing triggers the 25% tariff on every shipment.
The US Department of Commerce's Bureau of Industry and Security (BIS) announced Tuesday that Nvidia H200 chip sales to China would be permitted under specific conditions, effectively reversing the blanket export ban implemented by the Biden administration just one year ago.
The approval, which takes effect January 16, represents the culmination of a sustained industry advocacy effort that included direct engagement between Nvidia CEO Jensen Huang and President Donald Trump.
From restriction to approval: Nvidia's advocacy timeline
The policy shift follows a months-long campaign by Nvidia to convince US officials that overly restrictive export controls would simply surrender market share to competitors without meaningfully impeding China's AI development ambitions.
In January 2025, the outgoing Biden administration placed the H200 under export restrictions as part of the AI Diffusion Rule, threatening to cut Nvidia off from significant revenue in the world's second-largest economy as Chinese firms ramped up AI infrastructure investments.
By July 2025, Nvidia's advocacy efforts began showing results. Huang announced that the US government had promised to grant licenses for the company to resume exporting the H20 chip—a less powerful processor specifically designed for the Chinese market.
The H20 sales resumption signaled the Trump administration's willingness to reconsider the Biden-era restrictions.
The breakthrough came in December 2025, when Huang met directly with President Trump in Washington to discuss export controls. Days later, Trump publicly stated he would allow Nvidia to sell the H200 to approved customers in China, a clear indication that industry lobbying had reached the highest levels of government.
The conditions: Volume caps and verification requirements
The BIS approval stipulates that Nvidia H200 chip shipments to China cannot exceed 50% of the volume sold domestically in the United States. Additionally, every China-bound shipment must undergo verification by a US-headquartered third-party laboratory to confirm the processors' technical capabilities and functions.
The Trump administration on January 14 mandated that China-bound chips make a detour from Taiwan, where they are manufactured, through the United States for this third-party testing. When the chips enter US territory for verification, they become subject to the newly announced 25% tariff.
The White House clarifies the tariff won't apply to chips for US data centers, startups, consumer applications, or public sector use. However, chips transiting the US for testing before export to China fall within the tariff's scope, though Commerce Secretary Howard Lutnick retains broad exemption authority.
The licensing framework shifts from a presumption of denial to case-by-case review, giving Nvidia more flexibility than under the Biden-era blanket ban but maintaining government oversight of each transaction.
Notably absent from the final BIS announcement was Trump's December suggestion that 25% of chip sales would be "paid" to the US government—a proposal that raised questions about implementation feasibility and appears to have been dropped during policy formulation.
Market demand versus political headwinds
Nvidia's lobbying success comes at a time when the company faces strong demand signals from Chinese customers but growing political resistance from Beijing.
During last week's CES technology conference in Las Vegas, Huang described customer demand from China for H200 chips as "very high" and said Nvidia was "firing up" its supply chain in anticipation of resumed sales. The comments reflected the company's confidence that US approval would unlock significant revenue opportunities.
Chinese authorities have created their own obstacles. After Nvidia announced H20 sales resumption in July 2025, state media labeled the chip "unsafe" and urged boycotts, while the Cyberspace Administration summoned Nvidia to discuss alleged tracking and remote-control risks.
According to a report by The Information, the Chinese government has informed technology companies that it will approve H200 chip purchases only under exceptional circumstances, such as for university research and development laboratories.
The restriction suggests that while Nvidia won the lobbying battle in Washington, it may face continued resistance in Beijing as Chinese authorities push for domestic semiconductor self-sufficiency.
The 25% tariff adds another layer of complexity. With China already skeptical and the chips now carrying a significant price premium due to import duties, Nvidia faces the prospect of having secured legal permission to sell into a market that may prove economically unviable.
Nvidia shares traded slightly lower in after-hours trading following the tariff announcement, and the company has not yet commented publicly on how it plans to navigate the conflicting policy signals.
Industry implications and broader tech policy debates
Nvidia's successful advocacy campaign highlights the influence American technology companies wield over export control policy, particularly when they can argue that restrictions harm US competitiveness without achieving stated national security objectives.
However, Wednesday's tariff announcement reveals the contradictory nature of the Trump administration's semiconductor policy approach. The administration approved China exports one day, then imposed tariffs that could render those exports economically unfeasible the next.
The tariff follows a nine-month Section 232 investigation aimed at boosting domestic production. The US currently manufactures only 10% of required chips, creating what officials describe as a "significant economic and national security risk," with broader semiconductor tariffs possible in the near future.
The semiconductor industry counters that overly restrictive controls accelerate China's domestic development by creating guaranteed demand for local alternatives, arguing that maintaining some US market access allows American firms to fund innovation while monitoring Chinese technological progress.
National security advocates counter that providing China access to cutting-edge AI chips—even with restrictions—risks enabling military and surveillance applications that threaten US interests. The 50% volume cap in the H200 approval appears designed to address these concerns while giving Nvidia limited market access.
For technology companies across the Asia-Pacific region, Nvidia's lobbying success demonstrates that US export policy remains subject to industry influence and political negotiation rather than fixed strategic doctrine. The precedent may encourage other American technology firms facing export restrictions to pursue similar advocacy campaigns.
The BIS export policy takes effect January 16, 2026, with the tariff effective immediately. Nvidia now faces an unprecedented scenario: having lobbied successfully for market access only to see it potentially undermined by tariffs imposed 24 hours later.