World leaders want American artificial intelligence. They just don’t want America to be able to stop him.

At the G7 summit on Wednesday, world leaders such as French President Emmanuel Macron and Indian Prime Minister Narendra Modi expressed concerns that the United States could deny their countries access to the best American AI models at any time.

Macron warned G7 leaders and senior AI executives — including Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and President Donald Trump — over lunch that if the United States could “from one day to the next to flip off the switch,” it would not only damage the economies of European clients but also the AI ​​companies themselves.

These comments come just days after the Trump administration banned Anthropic from exporting its latest Mythos 5 and Fable 5 models on national security grounds. This order came after Amazon informed the White House that some safety barriers could be bypassed. Although cybersecurity experts have claimed that the capabilities mentioned by the government are also present in models that remain freely available, including OpenAI, Anthropic’s models remain on ice.

The incident exposed a risk that many international companies have been grappling with: any company or government that relies on U.S. AI infrastructure now has to consider the possibility that access will be revoked overnight, for reasons it may never be told about.

Prime Minister Modi also said he was concerned about Trump’s move to block the Anthropic Model, according to reports from the Financial Times, adding that democratic nations should have unfettered access to the best AI models to protect critical infrastructure.

“The recent restrictions on access to Anthropic models confirm what we at Cohere have known all along: that remaining companies and democratic nations dependent on a small handful of big tech companies poses a risk to resilience,” Aidan Gomez, co-founder and CEO of Canadian AI startup Cohere, said in a statement shared with TechCrunch. “Digital sovereignty is not just about competition in the marketplace or any one company or country. It is about who controls the underlying technology that will shape our economic security and national sovereignty for decades to come.”

During the meeting, G7 leaders also discussed creating a “Trusted Partners” plan that would give non-US countries access to advanced AI models from companies like Anthropic and OpenAI. The goal is to maintain some kind of open trade network that bypasses US restrictions. Both countries and companies could become reliable partners, as long as they use the models to develop stronger defenses against competitors like China.

But it’s not clear how far the trusted partner scheme will extend, or whether this is a solution for a startup in Paris or Bangalore that just had its product discontinued without warning.

Regardless, Macron noted that it would make sense for Washington to support such a scheme and ensure that Mythos is granted broader access. No one would want to buy access to American AI if it could disappear overnight.

These comments were made even as Europe and countries other than the United States try to push for AI supremacy — an issue that becomes increasingly difficult when American models continue to advance and no one wants to be left out.