In a move that could signal a pivotal shift in the global semiconductor landscape, President Donald Trump voiced his willingness to allow Nvidia—the world’s leading manufacturer of artificial intelligence (AI) chips—to sell a scaled-down version of its most advanced processor, the Blackwell, to China.

The announcement comes at a time when the intersection of national security, technology innovation, and global commerce is more fraught than ever before. Trump’s comments, made during a recent press conference at the White House, reflect both the growing pressures facing U.S. tech giants shut out of China and the evolving strategy of the Biden and Trump administrations to restrict China’s access to cutting-edge American technology.

What Did Trump Propose?

Trump indicated that he’d consider approving deals for Nvidia to export a downgraded version of the Blackwell chip—likely by cutting its maximum performance by 30% to 50%. “It’s possible I’d make a deal,” Trump stated, noting that Nvidia CEO Jensen Huang would soon meet with him to discuss the specifics.

He emphasized the importance of retaining America’s technological edge: “The Blackwell is the latest and greatest in AI chips, and we need to be careful about how much advantage we give away.”

Why Does This Matter?

The Blackwell architecture represents the pinnacle of AI computing power and serves as the backbone for training massive models used in fields from language processing to autonomous vehicles. U.S. export controls, which began with restrictions on Nvidia’s A100 and H100 chips, have intensified since 2022. China, accounting for up to 25% of Nvidia’s data center revenue before 2023, has scrambled to stockpile advanced processors as the U.S. tightened restrictions. The latest round of proposed deals would allow exports—but at the cost of performance, and with the U.S. government taking a reported 15% revenue cut from chip sales to China.

According to Paul Triolo, a senior vice president at the DGA-Albright Stonebridge Group, the aim is to “addict China to substandard, or non-cutting edge technology” and thus maintain America’s AI primacy.

China’s AI Push and Huawei’s Challenge

Following successive U.S. sanctions, China’s largest tech firms—including Alibaba, Tencent, and ByteDance—have relied on stockpiled Nvidia GPUs for AI development. However, with supplies dwindling and U.S. rules tightening, domestic champion Huawei has advanced its own Ascend series of AI chips. The Ascend 910B processor, launched in late 2024, is widely deployed in China’s data centers but still lags behind Nvidia’s top-tier offerings in both efficiency and broad developer ecosystem support.

Industry observers note a transition period: “Chinese firms are running out of their earlier-acquired Nvidia GPUs and betting on Huawei’s next-generation chips to bridge the gap,” says Triolo. Market analysts anticipate a more competitive chip from Huawei, possibly in 2026, but U.S.-developed chips remain the gold standard.

Regulatory and Geopolitical Context

Since 2022, successive U.S. administrations have layered export controls under the Commerce Department’s Bureau of Industry and Security (BIS) on sales of high-end semiconductors to China, citing concerns that the chips are used for military, surveillance, and intelligence purposes. The U.S. has also pressured allies like Japan and the Netherlands—home to lithography giant ASML—to restrict key technology exports to Chinese firms.

However, these measures come at a steep cost for American companies. Nvidia and AMD, which collectively dominate the market for processors used in data centers and AI research, risk losing billions in annual revenue. In 2023 alone, Nvidia’s data center sales reached nearly $50 billion, with China contributing a substantial portion before the harshest export curbs began.

The arrangement discussed—where Nvidia and AMD would pay a 15% cut of Chinese chip revenue to the U.S. government for licenses—reflects a shift toward regulatory pragmatism. While initial proposals by Trump’s administration sought a 20% cut, pressure from Nvidia led to the reduced figure and the potential for hardware downgrades as a compromise. The move is being closely watched by the global investor community and semiconductor industry.

What’s at Stake for Nvidia, AMD, and the U.S.?

For Nvidia, the world’s most valuable semiconductor company with a market capitalization over $2.8 trillion in Q3 2025, regaining even limited access to the Chinese market could add billions in offset revenues as worldwide demand for AI accelerates. CEO Jensen Huang has persistently argued that keeping Chinese firms “dependent” on American platforms will benefit U.S. interests in the long run—by preventing the rise of fully homegrown Chinese alternatives. Competitor AMD, a key player in both general-purpose and AI-optimized chips, is positioned similarly.

Critically, the U.S. government faces a delicate balancing act. Allowing chip sales bolsters domestic corporate earnings and affirms U.S. technological leadership, but risks arming a top strategic rival with advanced computational tools vital for military and economic prowess.

Global Implications and the Road Ahead

Should Trump’s proposal go forward, it would represent one of the most significant policy pivots since AI and semiconductor curbs began. Downgraded Blackwell chips—potentially capped in interconnection speed, processing power, or memory bandwidth—would likely still outclass many alternatives available to Chinese firms but ensure the U.S. maintains the latest edge.

The move could also pressure European and Asian chipmakers to calibrate their own export policies, as global supply chains remain acutely sensitive to regulatory changes. Investors in Nvidia, AMD, ASML, and related supply chain companies will be watching closely, as any deal would have reverberations across equity markets, trade flows, and the race for AI supremacy.

Meanwhile, China’s “AI self-reliance” campaign is gaining momentum, with government subsidies flowing to Huawei and other homegrown chip-makers. The coming years will see an intensified contest between export-controlled U.S. technology and indigenous innovation in China, with the outcome holding profound implications for the future of AI infrastructure worldwide.