Anthropic Disables Top-Tier AI Models After US National Security Order

The United States government has ordered Anthropic to disable its most advanced AI models — Fable 5 and Mythos 5 — for all foreign nationals, marking a dramatic escalation in how Washington controls frontier AI technology. Anthropic complied by shutting off access globally, not just for foreign users, in what experts are calling the first direct government-mandated takedown of a commercially deployed AI model.

The order came from the US Commerce Department, which cited a potential jailbreak that could allow Fable 5 and Mythos 5 to autonomously identify software vulnerabilities. Anthropic pushed back, stating it received only "verbal evidence of a potential narrow, non-universal jailbreak" and warned that applying this standard industry-wide would "essentially halt all new model deployments for all frontier model providers." The action moves US export controls decisively beyond chips and tools — now directly targeting the models themselves.

For Singapore, this is not distant geopolitics. The city-state's financial sector — from DBS to OCBC — runs on models like Claude for compliance, fraud detection, and customer service automation. Singapore's AI startups building on Anthropic's API layer now face sudden uncertainty about model availability. The government's own AI adoption push, from Smart Nation to the recently announced AI agent registry, depends on access to frontier models that may be subject to similar restrictions at any time.

The timing compounds the uncertainty. This comes just as Anthropic had confidentially filed for a US IPO, edging ahead of OpenAI. The relationship between Anthropic and the US government had already frayed earlier this year when the company refused to let the US military use its AI for domestic surveillance and autonomous weapons — leading to a supply chain blacklist set to take effect later in 2026. Pentagon CIO Kirsten Davies defended the directive, saying "some things are simply more important than revenue cycles."

Why it matters for Singapore: This shifts the calculus for every Singapore enterprise and government agency relying on US frontier AI models. It's no longer just about cost or capability — it's about geopolitical continuity. The move reinforces the case for Singapore's sovereign AI strategy and local model development, and it raises urgent questions about whether Singapore can count on US frontier models as a stable foundation for its digital economy.