Trump export ban on Anthropic rattles AI industry, allies

The Trump administration imposed sweeping export controls on Anthropic's AI models on June 12, forcing the San Francisco company to suspend global access just days after releasing its most advanced Claude upgrades—Fable 5 and Mythos 5.

The U.S. Commerce Department cited national security concerns around "jailbreaking," techniques that bypass AI safety guardrails. Anthropic countered that such vulnerabilities are minor and exist across competing platforms. The ban effectively blocks foreign nationals worldwide from using the models, including Anthropic's own international employees.

The timing proved particularly awkward: Anthropic had just filed plans for a fall IPO expected to raise tens of billions in capital.

Why it matters

Washington's willingness to deploy export controls against a leading American AI company—rather than reserving such measures for foreign adversaries—signals a new regulatory posture that could reshape investor confidence and international AI partnerships. The move has prompted allies to accelerate plans for independent AI infrastructure and question reliance on U.S. technology providers.

Tech sector and allies push back

More than 170 technology executives signed an open letter warning the restrictions "risked America's AI leadership" by denying cybersecurity defenders access to cutting-edge tools while China advances its capabilities.

At this week's G7 summit in Evian, France, the ban dominated side conversations. Indian Prime Minister Narendra Modi called for "broad and inclusive" access to U.S. AI models. The UK requested an exemption and was denied. French President Emmanuel Macron issued a blunt warning: "We won't buy any models made by these companies if overnight, you can just flip the switch."

European lawmakers described the episode as evidence of a U.S. "kill switch" over critical technology, reinforcing calls for the European Union to develop sovereign AI capabilities.

Regulatory vacuum fuels ad hoc decisions

Risto Uuk, head of European policy at the Future of Life Institute, called the move "hasty and uninformed," arguing that AI safety "cannot depend on a single firm's goodwill on a given week." He urged Washington to establish clearer regulations comparable to the EU's framework launching in August.

Clemens Fuest, president of Germany's ifo Institute, noted in a Friday research note that Europe controls less than 5% of global AI infrastructure, compared to 75% for the U.S. and 15% for China. He called the ban a wake-up call highlighting Europe's "vulnerability" and the need to expand data centers, chip manufacturing, and energy capacity within the bloc.

Pentagon tensions preceded the ban

Anthropic clashed with the Department of Defense earlier this year after refusing to remove restrictions preventing its models from enabling mass surveillance of U.S. citizens or fully autonomous weapons. The Pentagon designated Anthropic a "supply chain risk" and threatened to cancel contracts worth hundreds of millions, prompting legal action from the company.

Anthropic has positioned itself as a safety-first player in the AI race, using an approach called Constitutional AI that trains models to self-critique against written ethical principles.

IPO outlook remains uncertain

Bankers estimate Anthropic could raise $30 billion to $60 billion in its planned fall IPO, potentially one of the largest public offerings ever. The company was recently valued near $1 trillion and generates an estimated $47 billion in annual revenue.

Yet the export ban, Pentagon disputes, and lost government contracts create headwinds. IPO expert Jay Ritter of the University of Florida noted that prediction markets still show an 85% probability of an Anthropic IPO announcement before November 1—a figure largely unchanged since the ban. "There is still enormous enthusiasm in both public and private markets for AI companies," Ritter said.

At the G7 summit, Anthropic CEO Dario Amodei urged leaders to prioritize international cooperation on AI regulation over unilateral action, warning them to "resist the temptation to splinter."

These details were first reported by DW.