The US government now decides who gets access to frontier AI. Your AI roadmap just became a national security matter.

The Trump administration just told OpenAI to delay GPT-5.6.

Not a suggestion. A request backed by an executive order giving government cybersecurity teams 30 days to review advanced AI models before they go public.

Sam Altman told staff Wednesday: GPT-5.6 launches in limited preview only. Select enterprise customers. Government approves each one, case by case.

This is no longer a vendor relationship. It is a clearance process.

The June 2 executive order established a "voluntary" review framework. That framework just became operational reality for every enterprise betting on OpenAI's next model.

Anthropic got it worse. The administration issued an export control directive blocking Mythos 5 and Fable 5 access to foreign nationals, including Anthropic's own non-US employees.

Two labs. Two different levels of government control. Zero predictability for your procurement timeline.

If your enterprise AI strategy depends on GPT-5.6 or Claude Mythos, you are now subject to federal approval timelines you cannot negotiate, cannot expedite, and cannot predict.

Audit your model dependency today. Map which frontier models power your production workloads. Build contingency plans for 30-60 day delays on every major model release. Your AI roadmap is no longer a technology decision. It is a regulatory compliance exercise.

SOURCE: https://www.theverge.com/ai-artificial-intelligence/957372/openai-will-delay-gpt-5-6-after-trump-administration-request
VERIFIED: The Verge, The Information, Crypto Briefing, CyberSecurityNews
SIGNAL: Government now has effective veto power over who accesses frontier AI models. Every enterprise relying on OpenAI or Anthropic must factor regulatory delay into deployment timelines.