All topicsSTATE OF AI REPORT.

Can the US restrict access to frontier AI models abroad?

The June 2026 export directive required Anthropic to restrict Fable 5 and Mythos 5 access for foreign nationals, including those inside the US. Anthropic suspended access for all users because it could not immediately verify nationality. Its July 1 update announced restored access. Users outside the US lost access under a US policy decision.

Source: Anthropic: redeploying Fable 5.

Evidence you can use

The model-access restrictions and restoration

2026-06-12 to 2026-07-01

The model-access restrictions and restoration
DateEvent
2026-06-12SourceExport controls applied. Anthropic suspended both models for all users.
2026-06-30SourceAnthropic said the export controls had been lifted.
2026-07-01SourceAnthropic updated its notice to say access was restored.

Timeline based on Anthropic’s account. Restored access does not imply identical availability across every product, cloud, or customer. The company described separate arrangements for Mythos and its Glasswing partners.

Sources: Anthropic: redeploying Fable 5.

Europe’s dependence on US frontier models

As I argued in Europe cannot rent its way to AI sovereignty, a domestic data center does not confer control over the models inside it. A credible alternative requires sustained spending on researchers, compute, energy, and successive model generations. Europe's practical goal should be enough domestic capability to keep essential work running through a cutoff and negotiate with a bargaining chip in hand, which could be powered land and data center capacity. That is, if its energy prices and planning permission challenges can come down to be competitive…

Frequently asked questions

Answers drawn from the report and the sources below.

Sources and dates

2026 report snapshot. Preview revised 2026-10-07. Individual data periods and source checks are listed below. This is not a claim that every source was updated on that date.

  1. Anthropic: statement on the export directiveJune 2026. Primary source checked 2026-10-07.
  2. Anthropic: redeploying Fable 5Updated 2026-07-01. Primary source checked 2026-10-07.
  3. Nathan Benaich: Europe cannot rent its way to AI sovereignty2026. Author analysis.
  4. State of AI Report 2026, slide 166: Anthropic vs. US Government: who defines the limits of AI usage in defense2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  5. State of AI Report 2026, slide 167: Frontier AI goes live in US military operations2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  6. State of AI Report 2026, slide 168: Iran turned US commercial cloud infrastructure into an explicit military target set2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  7. State of AI Report 2026, slide 169: Outside the US and China, 67 countries have sovereign AI projects2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  8. State of AI Report 2026, slide 170: Selected sovereign AI program pledges total about $138B2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  9. State of AI Report 2026, slide 172: Korea is going big on funding domestic AI and building a market for it2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  10. State of AI Report 2026, slide 173: Governments are funding compute access for domestic AI developers2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  11. State of AI Report 2026, slide 177: Europe’s data center ambition is hampered by significantly more expensive energy costs2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  12. State of AI Report 2026, slide 178: China uses cheap power to favor domestic AI chips2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  13. State of AI Report 2026, slide 183: US states keep regulating AI despite Trump’s push for national rules2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  14. State of AI Report 2026, slide 184: California builds independent oversight of AI safety claims2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  15. State of AI Report 2026, slide 185: Brussels delays high-risk EU AI Act rules by up to 16 months2026 report snapshot. Read against the report PDF on 2026-10-07. Study-specific limits retained.
  16. State of AI Report 20262026 report snapshot. Report PDF. See individual slide references for the expanded analysis.
  17. CNAS: Sovereign AI Index2026 report snapshot. Retained from the launch essay.

Cite this page

Benaich, Nathan. “Can the US restrict access to frontier AI models abroad?.” State of AI Report 2026. Published 2026-10-08.