The White House says it wants America to win the global AI race. But a sudden move to block access to Anthropic's advanced models is doing the opposite, signaling to foreign governments that betting on U.S. artificial intelligence is risky.
The administration placed Anthropic's Mythos 5 and Fable 5 under export controls on Friday, effectively cutting off international customers who were already using these systems. Critics argue this amounts to a de facto licensing regime, coming just weeks after the White House explicitly rejected mandatory government licensing in a scaled-back executive order.
The timing has spooked America's allies. Canadian Prime Minister Mark Carney warned Sunday that relying on a single AI provider leaves nations vulnerable. European Union officials launched a "tech sovereignty" initiative this month aimed at reducing dependence on American technology companies, with plans to dramatically expand their own data center and semiconductor capacity.
Europe's message was blunt. "Europe wants to be in the position to make its own choices, avoiding risky dependencies on single dominant suppliers," said European Commission Executive Vice-President Henna Virkkunen.
What makes the Anthropic move particularly risky is the precedent it sets. Gartner noted Monday that this marks the first time a government has blocked access to an AI model that customers were already actively using, and warned such interventions will likely multiply. The technology research firm cautioned that "operational risk can stem not just from a vendor's performance but also from unpredictable government interventions."
The fear among foreign companies and governments is straightforward: if Washington can cut them off today, they need backup options. Chinese AI models, though roughly six months behind U.S. capabilities, are increasingly attractive as insurance policies, especially open-sourced versions that pose minimal risk to adopt. For certain applications, they may prove compelling alternatives precisely because they come without American strings attached.
The White House pushed back on criticism, with spokesman Kush Desai saying the administration is "collaborating with AI industry leaders to balance cutting-edge innovation with national security concerns." He reiterated that the U.S. remains "by far the world leader in the global AI race."
That advantage, however, is real only if allies actually use American technology. The deeper problem is Washington's contradictory signals. In recent weeks, Trump delayed an executive order on voluntary AI reporting to avoid threatening America's lead. Then it issued a slimmed-down order rejecting mandatory licensing. Then it weaponized export controls to achieve what it said it wouldn't do.
Until the administration settles on a coherent approach, foreign governments will keep hedging their bets, and U.S. AI makers will keep losing customers who can't afford unpredictable cutoffs.
Author James Rodriguez: "The Anthropic crackdown reveals a fundamental contradiction at the heart of Trump's AI strategy: you can't threaten to pull the plug on allies and expect them to stay dependent on you."
Comments