A major player in the artificial intelligence market, Anthropic, has entered direct negotiations with the White House administration. The goal is to lift the imposed export restrictions that have effectively paralyzed the operation of two flagship AI models: Fable 5 and Mythos 5. These models were disabled for users outside the U.S. just days after their launch.

The conflict escalated after an executive order was signed on Friday, restricting foreign citizens' access to advanced AI developments. In response, Anthropic was forced to urgently block its newest products for an international audience.

By Saturday, the first round of intensive negotiations had taken place. The parties resumed dialogue almost immediately. On the government side, Commerce Secretary Howard Lutnick and Cybersecurity Director Sean Cairncross participated in the discussions. Anthropic was represented by co-founder and head of computing Tom Brown, as well as policy director Sarah Hack. Company engineers also held a series of online meetings with White House representatives.

In Search of a Compromise

Both sides are interested in resuming the models' operation as soon as possible, but a specific mechanism for a solution has not yet been determined. Within government circles, a scenario is being considered where Anthropic engineers would join forces with specialized security experts to develop mutually acceptable terms. This could set a precedent for the entire industry, establishing a new standard for interaction between AI developers and regulators.

The situation extends far beyond a typical discussion of artificial intelligence policy. Anthropic recently confidentially filed an S-1 application with the SEC, which is the first step toward preparing an initial public offering (IPO). A prolonged conflict with authorities could negatively impact the company's perception among potential investors and complicate the process of going public.

Expert comment: This case clearly demonstrates how fragile the balance between national security and technological development can be. For Anthropic, on the verge of an IPO, every day of downtime for its flagship models represents not only lost profits but also a serious reputational risk. The outcome of these negotiations will set the tone for AI regulation in the U.S. for years to come.