Trump rejects AI slowdown, doubles down on China race narrative

Trump’s stance crystallizes the political fault line that has been building for months. On one side are executives at Anthropic, OpenAI and xAI, now openly flirting with the idea of coordinated slowdowns to keep frontier systems manageable. On the other are nationalist arguments that any deliberate pacing hands China an opening and is therefore unacceptable. By siding emphatically with speed, Trump signals that, at least in the US, AI safety may get subordinated to geopolitical competition whenever the two appear to conflict.

This puts labs in an uncomfortable position. They are simultaneously warning of serious systemic risks and being told by the country’s leadership that slowing down would be unpatriotic. That tension could push companies to treat “slowdown” proposals as negotiation chips to extract liability protections or subsidies rather than as genuine commitments.

For AGI, tying model development to national prestige and great power rivalry tends to shorten time horizons. It encourages labs and their backers to prioritize capability milestones and market dominance, assuming safety work can catch up later. History suggests that when those instincts collide with calls for restraint, speed usually wins unless there is an unusually strong countervailing institutional check.