The US killed Fable 5. Four open models filled the gap in 6 days. Your vendor strategy is now a compliance risk.

The Commerce Department sent Anthropic a letter at 5:21 PM on June 12.

The order: disable Fable 5 and Mythos 5 for every foreign national on Earth.

Including Anthropic's own foreign-national employees.

Anthropic couldn't filter users by nationality in real time across dozens of cloud platforms.

So they killed it for everyone. AWS Bedrock. Google Cloud. Microsoft Foundry. Snowflake. The direct API. All of it. Gone in an afternoon.

If you built automation on Fable 5, you lost your engine on a Friday evening.

But here's what happened next that your CISO missed.

Between June 9 and June 13, four open-weight models became enterprises' instant second source.

Cohere shipped North Mini Code. 30B parameters, 3B active per token. Apache 2.0 license. Runs on a single H100. A bank can self-host it without a legal review.

Moonshot shipped Kimi K2.7-Code. 1 trillion parameters. 95 cents per million input tokens. That's a fraction of hosted frontier pricing.

Zhipu opened GLM 5.2. One-million-token context window. Compatible with Claude Code, Cline, OpenCode out of the box.

Cohere's chief AI officer told Bloomberg they're seeing a "huge number of inbounds" from enterprises diversifying away from US providers. Governments outside the US and China are scrambling.

Trump told Axios on June 19 he no longer views Anthropic as a security threat. But Fable 5 is still offline. No formal restoration order exists.

Audit your model dependencies today. If your production systems depend on a single frontier vendor, the US government just proved it can kill your AI stack with one letter.

Build a multi-vendor strategy with at least one sovereign or open-weight fallback. This is not optional anymore. It is infrastructure resilience.

SOURCE: https://www.anthropic.com/news/fable-mythos-access
VERIFIED: Anthropic official statement (June 12, 2026), The New Stack (June 18, 2026), BetaKit/Bloomberg (June 16, 2026)
SIGNAL: The first government-forced takedown of a deployed frontier AI model just rewrote every enterprise vendor risk framework. Single-vendor AI is now a national security liability.
---