Breaking

Trump Tells Anthropic, OpenAI and xAI He Will Not Slow Frontier AI Over China

September 13, 2026 at 17:50 EDT

Policy & Regulation
Foundation Models
AI Agents

President Donald Trump on Sunday rejected a public appeal from the chief executives of Anthropic, OpenAI, and xAI to slow the development of more powerful AI models, arguing that the United States cannot afford to lose its lead over China.

September 13, 2026 · Doonbeg, Ireland

No brake on the frontier

President Trump rejected a rare call from Anthropic, OpenAI, and xAI to slow frontier AI, saying the United States must keep its lead over China because whoever wins AI, wins.

Time on two clocks

Months to feared swarm scale vs extra time labs want

6 mo

Swarm risk, low

12 mo

Swarm risk, high

24 mo

Extra time, high

Each block is six months. Amodei warned a misaligned agent swarm could scale in 6–12 months; he said an extra year or two could buy safer alignment.

Pacing steps, not a halt

~1 yr

U.S. lead over China, per Trump

30 vs 90

Days of voluntary pre-release review (June order vs earlier draft)

64%

Expect AI to cut U.S. jobs (CBS poll)

Amodei’s plan

1 · Embedded evaluators

Outsiders with desks, badges, laptops inside the lab.

→

2 · Democratic coordination

Shared standards and limits among labs in democracies.

→

3 · Global coordination

A China deal, if verification can be made ironclad.

Keep building

China will not wait

Trump and allies say a pause would let Beijing overlap the United States. Labs that want to go slower, they argue, can do so without a government mandate.

Pace the frontier

Safety is lagging capability

Amodei, Altman, and Musk backed slower capability gains. Critics say even that is too weak, and some want a moratorium on self-improving systems.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.