Nvidia CEO Jensen Huang Discusses Chip Controls with Trump, Slams Regulation

Nvidia CEO Jensen Huang met with former President Donald Trump on Wednesday to discuss the U.S. government’s proposed export restrictions on advanced artificial‑intelligence chips. The conversation, held on Capitol Hill, centered on a pending legislative effort that would limit the sale of cutting‑edge AI processors to certain foreign markets, most notably China.

Huang reiterated Nvidia’s support for export controls, emphasizing that “American companies should have the best, the most, and the first” when it comes to advanced semiconductor technology. He added that the industry’s leadership role in AI hardware should not be compromised by premature technology transfers.

The legislation in question is the Guaranteeing Access and Innovation for National Artificial Intelligence Act (GAIN AI Act), which lawmakers have considered inserting into the National Defense Authorization Act (NDAA). The bill would require U.S. chipmakers, including Nvidia and Advanced Micro Devices (AMD), to give domestic firms priority access to their most sophisticated AI chips before exporting them abroad.

According to industry sources, the GAIN AI Act is unlikely to be attached to the final NDAA package. Huang welcomed the omission, calling the proposal “more detrimental to the United States than the AI Diffusion Act” and warning that it could stifle innovation and erode the competitive edge of U.S. firms.

Huang also warned against a fragmented, state‑by‑state approach to AI regulation. He argued that a patchwork of state laws would “drag this industry into a halt” and pose a national‑security risk by slowing the United States’ progress in AI research and deployment. Instead, he advocated for a single, federal regulatory framework that would streamline compliance and preserve the nation’s leadership in the sector.

Former President Trump has publicly urged Congress to include a provision that would pre‑empt state AI regulations in favor of one federal standard. However, House Majority Leader Steve Scalise (R‑LA) told reporters that the pre‑emptive language lacks sufficient support to make it into the defense bill. Scalise indicated that lawmakers will continue to explore ways to establish a national AI standard.

Business and Technical Implications

From a market perspective, tightening export controls could reshape the global AI hardware supply chain. Nvidia’s flagship GPUs—such as the Hopper and the upcoming Ada‑level architectures—are critical for training large‑scale language models and deep‑learning workloads. If the United States restricts shipments to China, Nvidia may see a short‑term dip in revenue from one of its fastest‑growing regions. However, the company could offset this loss by expanding sales to domestic cloud providers, defense contractors, and other allied nations that are increasing their AI compute budgets.

For AMD, which has been gaining ground with its MI series accelerators, the impact would be similar. Both firms rely on a sophisticated ecosystem of software stacks (CUDA, ROCm) and a deep talent pool in the United States. Limiting exports could accelerate domestic demand for AI‑optimized silicon, encouraging further investment in U.S. fabs and potentially spurring the construction of additional manufacturing capacity at facilities like TSMC’s Arizona plant.

On the technical front, the proposed restrictions target chips classified under the Export Administration Regulations (EAR) as “high‑performance computing” (HPC) items—typically those with an export control classification number (ECCN) of 3A001 or 3A002. These chips enable the training of models with billions of parameters, a capability that governments consider a dual‑use technology. Companies will need to implement rigorous compliance processes, including end‑use verification and end‑user licensing, to avoid inadvertent violations.

The broader geopolitical context is also at play. China’s push to achieve AI self‑sufficiency has led to heavy investment in domestic semiconductor initiatives, such as the “Made in China 2025” program and the recent launch of home‑grown AI accelerators. U.S. export controls could incentivize Chinese firms to accelerate their own chip development, potentially narrowing the technology gap over the long term. For U.S. firms, the strategic calculus involves balancing short‑term revenue considerations with the long‑term goal of preserving a technological moat.

Industry lobbying is intensifying. An AI‑focused super‑PAC, dubbed “Leading the Future,” has been established to rally support for a cohesive federal regulatory approach. The PAC’s fundraising efforts underscore the high stakes for semiconductor giants and AI startups alike, as regulatory certainty is essential for forecasting R&D spend and capital allocation.

In summary, while the immediate effect of the GAIN AI Act’s exclusion from the NDAA may relieve pressure on Nvidia and AMD, the underlying debate over export controls and AI governance remains unresolved. Companies will need to navigate a complex regulatory landscape, manage supply‑chain risks, and adapt their go‑to‑market strategies as Washington continues to weigh national‑security concerns against the desire to maintain U.S. leadership in artificial intelligence.