The Trump administration is said to be exploring government equity stakes in both OpenAI and Anthropic, with last week's suspension of Anthropic's Claude Fable 5 and Mythos 5 seen by some as a potential bargaining lever.
June 2026 · Government & AI
Washington Eyes Equity in AI Giants — As Anthropic's Newest Models Go Dark
The Trump administration is weighing government stakes in OpenAI and Anthropic. Days later, an export-control order forced Anthropic to fully shut down its just-launched Claude Fable 5 and Mythos 5 — seen by some as leverage in those very talks.
3 days
From launch (Jun 9) to full shutdown (Jun 12)
1st
Time export controls applied directly to an AI model itself
~200
Organizations with Mythos 5 access (Project Glasswing)
Benchmark: SWE-Bench Pro
The shut-down model topped the previous flagship on long-horizon software engineering.
The shut-down models at a glance
Context
1M tokens (128k output)
Knowledge cutoff
January 2026
Price (per M tokens)
$10 in / $50 out
How the shutdown happened
Commerce Dept. export-control order: cut off all foreign nationals
→
Filtering users by nationality proved effectively impossible
→
Anthropic shuts both models down entirely
At launch — praise
"Months of engineering compressed into days"
Rebuilding web apps from screenshots
Faster vulnerability fixes & drug discovery
After shutdown — concern
Fears of a precedent-setting "kill switch"
Green-card holders losing access
Devs advised to fall back to Opus 4.8
Neither the equity talks nor the model restrictions have reached a conclusion — leaving the relationship between the U.S. government and leading AI firms open to further negotiation.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…