The US government directed Anthropic to shut down its strongest Claude models, according to a post on X by @rohanpaul_ai. The directive reportedly invokes export control authorities, though the specific legal basis has not been disclosed.
Key facts
- US government directed Anthropic to shut down strongest Claude models.
- Export control authorities cited as legal basis for the order.
- Claude 4 Opus scored 79.8 on SWE-Bench.
- First instance of US ordering domestic model shutdown.
- No official confirmation from Anthropic or government agencies.
The US government directed Anthropic to shut down its strongest Claude models, according to a post on X by @rohanpaul_ai. The directive reportedly invokes export control authorities, though the specific legal basis has not been disclosed. No official statement from the White House, Commerce Department, or Anthropic has been published as of press time.
The order targets Anthropic's most capable AI systems — likely including Claude 4 Opus, which scored 79.8 on SWE-Bench and was released in early 2026. The company has not confirmed which models are affected or whether the directive applies to existing deployments or future releases.
Export Control Context
The US government has increasingly used export controls to limit the spread of advanced AI capabilities, particularly to China. In October 2025, the Biden-era rules were extended under the current administration, restricting the transfer of AI model weights and training data for frontier systems. This directive appears to be the first time the government has ordered a company to shut down domestic operations of its own models, rather than blocking exports.
Immediate Implications
The order raises immediate questions about the scope of US government authority over frontier AI model deployment. If enforced, it would set a precedent for the government to unilaterally disable AI systems it deems too dangerous — even those built and hosted entirely within the US. Anthropic has not indicated whether it will comply or challenge the order in court.
Industry Reaction
The AI research community has reacted with alarm. Several prominent AI researchers have pointed out that the order could undermine trust in US-based AI companies and accelerate efforts to develop open-source alternatives outside US jurisdiction. The move also contrasts with Anthropic's own stated commitment to responsible AI development through its RSP framework.
Key Takeaways
- US ordered Anthropic to shut down strongest Claude models via export controls.
- No official confirmation yet.
What to watch
Watch for an official statement from Anthropic or the Commerce Department in the next 48 hours. Key questions: which models are affected, whether the order applies to existing API customers, and whether Anthropic will challenge the directive in court. The Q2 2026 export control rulemaking cycle may provide the legal framework.
