G7 nations and multinationals rethink AI dependency after US export controls freeze Anthropic access
On June 12, 2026, a Commerce Department directive suspended all foreign-national access to Anthropic's newly released Fable 5 and Mythos 5 models, citing national security concerns. The freeze lasted days. The policy conversation it started may last years.
What happened
Commerce Secretary Howard Lutnick issued the directive at 5:21 pm ET on June 12, per AI News. It applied to any foreign national, inside or outside the US, including Anthropic's own employees who hold non-US citizenship. Anthropic disabled both models for all customers to ensure compliance. The White House cited a jailbreak method that circumvented Fable 5's safety controls, though Anthropic disputed the severity of the technique, saying the capability to "review programme code and identify errors" was common to rival models including OpenAI's GPT-5.5.
Compliance now requires individually validated licences from the Commerce Department's Bureau of Industry and Security, with civil penalties attached. The access restoration path, as David Sacks put it, falls squarely on Anthropic: "The ball is in Anthropic's court."
The international fallout
Within 48 hours, European politicians had reframed the episode as a structural dependency problem. Finnish MEP Aura Salla said Europe "cannot continue to increase its technical potential by relying on access that can be turned off by a foreign government overnight," per IAPP. Former French interior minister Bruno Retaillau called the shutdown "a wake-up call," saying "a nation that depends on others for its technology is a nation that can be unplugged overnight." British MP Al Carns noted that researchers, companies, and hospitals in the UK had been piloting the models and were cut off with no advance notice.
The reaction aligned with policy work already in motion. The European Commission had proposed the European Technological Sovereignty Package earlier this month, and Canada's National Artificial Intelligence Strategy, released last week, explicitly named resilience against overreliance on a small number of American providers as a strategic objective, per IAPP.
The enterprise risk that wasn't in the playbook
The more durable consequence may be inside corporate risk teams rather than government ministries. Baker McKenzie partner Brian Hengesbaugh, writing for IAPP, called the incident the latest "iteration of how geopolitical risk impacts multinational companies" and recommended that enterprises begin incorporating "potential unavailability of AI models into company workflows" as a business continuity planning item. That framing matters: until June 12, most enterprise AI deployment risk models focused on data privacy, model accuracy, and vendor lock-in. Access termination without advance notice and without a court order was not a mainstream scenario.
The precedent, as IAPP noted, is not limited to Anthropic. The export control mechanism can apply to any AI model with dual-use capabilities. A multinational that built a workflow around a US-hosted frontier model now holds a concentration risk that is difficult to hedge quickly.
What to watch next
Three things will signal how this plays out. First, whether the European Technological Sovereignty Package advances to binding legislative form or stays a framework document. Second, whether any G7 government announces a state-backed frontier model specifically to provide an alternative to US-controlled infrastructure. Third, how enterprises change AI procurement criteria in Q3 and Q4 vendor reviews. The G7 summit in the French Alps this week included AI sovereignty as an agenda item beneath the surface, per Tech Policy Press contributing editor Mark Scott.