US Commerce Department says Mythos 5 and Fable 5 ban was meant to keep the models from Chinese and Russian intelligence

US Commerce Secretary Howard Lutnick said the ban on foreign nationals accessing the Mythos 5 and Fable 5 AI models, which took effect on Friday, July 10, 2026, was prompted by the risk of the models falling into the hands of intelligence agencies in hostile countries, including Russia and China.

What was banned and when

The ban took effect on Friday, July 10, 2026, and from that point on, foreign nationals were blocked from accessing the Mythos 5 and Fable 5 models. The US Department of Commerce made the decision, and Lutnick publicly disclosed the department's reasoning for the first time only after the ban was already in force.

• The ban took effect on Friday, July 10, 2026
• The decision was made by the US Department of Commerce
• Commerce Secretary Howard Lutnick publicly explained the reasoning
• The stated reason: the risk of the models being used by Russian and Chinese intelligence
• The restriction covers the Mythos 5 and Fable 5 models

Why officials are worried about these models in particular

Lutnick tied the restriction to the possibility that the tools could end up with intelligence agencies of states hostile to the US. That framing puts Mythos 5 and Fable 5 in the same category as dual-use technologies: tools that are useful for civilian work but also attract military or intelligence interest. In recent years, this logic has increasingly shaped US export controls on frontier AI models and the chips used to train them.

How this fits into broader US export control policy

For the past several years, the US Department of Commerce has steadily expanded the list of technologies subject to export controls on national security grounds. It started with advanced AI training chips, then extended the restrictions to the equipment used to manufacture them. Now, judging by Lutnick's comments, the department is ready to apply the same logic to the models themselves — that is, to the software, not just the hardware it was trained on.

• Agency: US Department of Commerce (U.S. Department of Commerce)
• Stated risk: foreign intelligence access to frontier AI models
• AI training chips and manufacturing equipment have previously been subject to similar restrictions

How this case differs from earlier restrictions

Export controls on chips mainly affect hardware makers — companies like NVIDIA, which have to get approval for shipments to certain countries. Restricting access to specific AI models is a different level of regulation: the target is no longer physical equipment but a lab's software product, which could potentially be copied or distributed far more easily than a shipment of chips. That is why Lutnick's framing of an intelligence risk reads as a signal that regulators now treat the models themselves, and not just the compute used to train them, as an object of strategic control.

What it means

The commerce secretary's direct, public explanation of the ban shows that US authorities view frontier AI models not only as a commercial asset but also as an intelligence asset — and that they are willing to impose export controls on the same national security grounds previously applied to chips and weapons.