# Anthropic calls for a coordinated global slowdown on AI

> On 4 June 2026 Anthropic urged a coordinated, verifiable mechanism to slow frontier AI.

*A serious warning, an awkward week. The brief.*

By The FeaturedDaily Desk · FeaturedDaily
Canonical: https://featureddaily.com/news/anthropic-slowdown-call-brief

> **Key:** **The one-liner:** the lab building one of the most powerful models asked the world to build a way to slow such models down — the same week it raised compute and prepped an IPO.

**What happened.** On 4 June, Anthropic published *When AI Builds Itself*, calling for a coordinated, verifiable mechanism to slow or pause frontier AI before systems can meaningfully improve themselves. Its illustration: it says Claude now writes over 80% of the code merged into Anthropic's own codebase.

**Not going it alone.** Anthropic explicitly rejects a unilateral pause — that would just hand the lead to less-cautious labs — and frames this as a US–China collective-action problem. OpenAI's leadership echoed the idea on 8 June.

> **Note:** **The friction.** The call coincided with reporting of a ~$35bn financing platform for Anthropic compute and a coming IPO, prompting White House and researcher pushback that it's 'safety as a moat'.

## Key takeaways

- What: a call for a coordinated, verifiable mechanism to slow or pause frontier AI development.
- Why: AI is increasingly building AI — Anthropic says >80% of its merged code is Claude-written.
- Not unilateral: Anthropic won't stop alone; it wants the US and China to coordinate.
- The friction: OpenAI echoed it, but it landed amid a ~$35bn compute push and a planned IPO.

## FAQ

### What did Anthropic actually call for?
A coordinated, verifiable mechanism — across labs and countries — to slow or pause frontier AI when risks demand it, plus government power to block dangerous deployments. Not a unilateral, one-company stop.

### Why are people skeptical?
Because it arrived alongside a ~$35bn compute push and a planned IPO. Critics argue a coordinated freeze would lock in current leaders, so the warning may be sincere and self-serving at once.