Amazon's Andy Jassy warned US authorities about Fable 5 risks — Anthropic shut down models

CEO Amazon Andy Jassy became the source of warnings that prompted the US government to demand the immediate shutdown of Fable 5 and Mythos 5 — Anthropic's most powerful models — for all customers simultaneously. According to The Wall Street Journal, it was Jassy's conversation with federal officials that triggered the chain of events culminating in the forced recall on Friday.

What happened on Friday

Jassy held a private meeting with Treasury Secretary Scott Bessent and other senior officials. The Amazon head reported that researchers at the company used Fable 5 to obtain information that raised serious national security concerns. Details of the specific incident have not yet been disclosed publicly — it is unknown what data was obtained or whether the use was intentional. Following this conversation, authorities demanded that Anthropic act immediately. On Friday, the company complied: Fable 5 and Mythos 5 were simultaneously disabled for all customers — corporate and private — without any prior notification. Services running on these models stopped without warning.

Scale of consequences

This is an unprecedented case for the American AI industry. For the first time, a commercial company was forced to freeze flagship products by direct government order — not due to a technical failure, not by its own decision, but as a result of state intervention on security grounds.

• Fable 5 — Anthropic's most powerful model, designed for analytics, research and strategic tasks
• Mythos 5 — its extended version with additional capabilities for handling large data volumes
• Both models were widely used by corporate clients, research organizations and government agencies
• The shutdown occurred without warning and without the ability to switch to backup systems
• Amazon — Anthropic's largest corporate investor with total investments exceeding $4 billion

For clients who had built critical processes on these models, the shutdown was painful: no migration plan, no timeline, no alternative solutions were offered.

Conflict of interests

The situation appears structurally ambiguous. Amazon is simultaneously Anthropic's largest corporate investor and the primary distribution platform for its models — it is through Amazon Web Services that most corporate clients gain access to Claude. A paradox emerges: the top executive of the investor company directly informed federal authorities about risks associated with the product of the same company in which billions of dollars have been invested. Why Jassy chose to go through government channels rather than internal negotiations with Anthropic leadership or the board of directors remains unclear. It's possible he considered the risks too serious for a corporate solution.

"When an investor directly signals regulatory concerns about its

portfolio company — this sets a precedent that the AI industry has not yet seen."

AnthropIC has not officially commented on when the models might return to service or what conditions would enable this. It also remains unclear whether measures will be limited to specific use cases or whether access will remain blocked indefinitely.

What this means

The Fable 5 and Mythos 5 story sets a new level of regulatory risk for the entire AI industry. Now, a single conversation between a key industry player and officials is enough for a competitor's flagship product to be forcibly withdrawn without public hearings, without court proceedings and without customer notification. For companies building critical infrastructure on AI models, the question of provider diversification has shifted from theoretical to entirely practical.