Anthropic, the AI safety lab valued at nearly $1 trillion, has called for a coordinated global pause in frontier AI development.
In a proposal published June 4, 2026, through its Anthropic Institute, the company argues that AI systems are approaching the ability to recursively improve themselves, and humans are losing meaningful oversight.
The warning is backed by internal data: as of May 2026, over 80% of code merged into Anthropic’s own codebase was written by Claude, not human engineers. Engineers are merging eight times more code per day compared to 2024, and AI task-completion horizons have doubled roughly every four months.
Anthropic acknowledges the game theory problem-a unilateral pause would leave it vulnerable to competitors like OpenAI, Google DeepMind, and xAI. The solution: internationally verifiable mechanisms enabling simultaneous halts across labs.
Unlike the 2023 Future of Life Institute open letter, which was largely ignored, this call comes from inside a frontier lab with firsthand access to the capabilities it warns about.
Anthropic is reportedly preparing for an IPO, with its valuation between $965 billion and $1 trillion. A coordinated slowdown would freeze the competitive landscape, cementing the advantages of existing leaders and locking out newcomers.