OpenAI has gone quiet on keeping AI chips from China, and Nvidia is the big winner

OpenAI has gone quiet on keeping America’s best AI chips out of China, just as Congress gets ready for its last big fight over the issue this year. Capitol Hill aides and lobbyists told Politico that the ChatGPT maker, which publicly backed tough chip export controls for much of last year, has stepped back from pushing for new restrictions, leaving the field to Nvidia, the chipmaker that wants to sell more of its chips in China.

The timing matters. China hawks in Congress want to attach new chip limits to the annual defence bill during the lame-duck session after next month’s midterms, and they had been counting on OpenAI’s support, according to Politico. The shift also brings OpenAI closer to President Trump, who has allowed some Nvidia chip sales to China and has close ties to Nvidia’s boss, Jensen Huang.

“Silence is doing Nvidia’s work for it”

None of the people Politico spoke to were named, and OpenAI hasn’t said publicly that its position has changed. But one congressional aide described a clear cooling:

They were with us on this and now their position is a bit more fuzzy. Silence is doing Nvidia’s work for it.

A congressional aide, speaking to Politico

A lobbyist put it down to politics, telling Politico there is “no appetite to challenge the president at this moment”.

OpenAI’s statement to Politico didn’t address its meetings with lawmakers or officials. A spokesperson said that “US labs having a substantial lead in advanced chips is key” to the company’s goals, and called effective export control policies “one part of that broader strategy”. Nvidia declined to comment. A White House official disputed the idea that Trump had loosened the rules, calling the administration’s approach “the most rigorous export control regime in modern history”.

What OpenAI said before

OpenAI has put its chip views in writing before. In its March 2025 proposals for the White House’s AI Action Plan, it called for an export strategy that would “use export controls to protect America’s AI lead”, sorting countries into tiers, with China among those to be kept out. By June this year, according to Politico, a federal AI framework OpenAI proposed mentioned the subject in a single line, calling to “strengthen export controls” without naming China.

OpenAI has also been close to the White House lately: its president, Greg Brockman, was at last month’s lunch where Trump said AI bosses had signed a “morally binding” pact that comes with no new rules.

Anthropic went the other way

The contrast with Anthropic is sharp. In September, Anthropic quit the Information Technology Industry Council, a trade group whose members include OpenAI, Google and Nvidia, after it lobbied to strip three chip bills from the defence bill: the AI OVERWATCH Act, which would give Congress a review period before advanced chips are licensed to adversary nations; the Chip Security Act, which would make companies track where their AI chips end up; and the MATCH Act, which would push allies to restrict sales of chipmaking equipment to China. Anthropic backs all three. It is also under fire this week over a leaked draft of its own federal AI bill.

The fight is live on both sides of the Pacific. US rules still let Chinese firms rent American chips in overseas data centres, Beijing is weighing whether to let ByteDance and Alibaba buy Nvidia’s newest chip, and last week China proposed its own export controls on 23 AI technologies.

Why it matters

OpenAI is the best-known name in AI, and its voice carries weight with lawmakers deciding how many advanced chips China can get. If it stays quiet through the lame-duck session, the chip bills lose one of their most powerful backers just when they have their best chance of passing, and Nvidia’s argument for selling more abroad meets less resistance.

Sources: Politico; OpenAI, proposals for the US AI Action Plan (March 2025); Axios.