Jensen Huang Meets Trump on Chip Exports, Hits State AI Rules While OpenAI Deal Stalls
Nvidia CEO Jensen Huang grabbed headlines this week with a sit-down with President Trump on chip export limits and sharp words against fragmented AI regulations. Separately, the company’s CFO revealed that a massive planned investment with OpenAI—once hyped by Huang himself—still lacks a final contract two months after the announcement.
Huang-Trump Talk and Export Controls
Huang met President Donald Trump on Wednesday to discuss export controls in broad terms, CNBC reports, while Reuters notes Trump praised him. He reiterated Nvidia’s support for such measures to keep American firms ahead. This comes as Congress eyes AI chip sales curbs in a defense bill, though a specific proposal called the GAIN AI Act looks unlikely to make the cut, a win for Nvidia lobbying, Bloomberg says.
Huang voiced uncertainty whether China would accept Nvidia’s H200 chips even if U.S. restrictions ease, Bloomberg reports.
Huang called excluding the GAIN AI Act from the National Defense Authorization Act a smart move. He labeled it worse for U.S. interests than another bill, the AI Diffusion Act.
Push for Federal AI Rules Over State Patchwork
Huang slammed the risk of states crafting their own AI laws, saying it would halt industry progress and hurt national security by slowing U.S. AI advances. He argued for a single federal standard instead, positioning Nvidia as the key player, Axios reports.
Trump recently backed a national AI standard to override states, but House Majority Leader Steve Scalise said Tuesday it won’t reach the defense bill due to insufficient backing. Lawmakers plan other paths forward.
$100 Billion OpenAI Deal Not Signed Yet
Shifting to partnerships, Nvidia EVP and CFO Colette Kress said at the UBS Global Technology and AI Conference that the eye-popping OpenAI arrangement remains at the letter-of-intent stage, Fortune reports. No definitive agreement exists despite the September reveal of up to $100 billion in Nvidia investments tied to 10 gigawatts of data center capacity.
Huang had dubbed it “the biggest AI infrastructure project in history,” with analysts eyeing up to $500 billion in potential Nvidia revenue from millions of GPUs over years. Kress stressed the long partnership—OpenAI sees Nvidia as its top compute provider—but noted current sales forecasts exclude this deal. OpenAI buys now through Microsoft and Oracle, aiming to go direct once finalized.
- Nvidia’s 10-Q filing flags risks: no guarantee investments close, supply chain snags from pre-paid GPU orders, power shortages delaying data centers, and fast innovation cycles risking demand mismatches.
- Competition from chips like Google’s TPU? Kress said no—Nvidia’s edge lies in its full stack of hardware, CUDA software, and tools.
These updates show Huang steering Nvidia through policy fights and big bets amid AI’s buildout challenges.
