The world's most powerful AI companies just watched the White House blink first.
The Summary
- Trump scrapped a planned executive order requiring government safety reviews of new AI models hours before signing, citing competition with China as justification
- The reversal came despite growing public backlash and expert warnings about critical security risks from unrestricted AI model releases
- Big tech gets a clear signal: safety theater is optional when market dominance is at stake
The Signal
The executive order that almost was would have established pre-release government safety reviews for new AI models. Not audits. Not red-teaming by third parties. Full government review gates before launch. Hours before signing, Trump killed it, telling reporters the US wouldn't "do anything to get in the way" of tech firms competing with China.
This wasn't a policy debate. It was a capitulation. The China card works every time because it short-circuits every other consideration. National security risks? Irrelevant if China ships first. Public concern about AI safety? Noise compared to the threat of falling behind Beijing. The playbook is clean: invoke the competition frame, watch regulators scatter.
"Trump vowed the US government would not slow down the AI race."
What's remarkable is the timing. The order was written, vetted, scheduled. Then someone made a call. We don't know who, but we know the argument: any friction on US AI development is a gift to China. Whether that's true is beside the point. It's unfalsifiable and it works.
Here's what this means in practice:
- OpenAI, Anthropic, Google, and Meta can ship foundation models without pre-release government safety checks
- "Self-regulation" remains the industry standard, which is to say no standard at all
- The safety review infrastructure that would have built institutional knowledge about AI risks won't get built
The gap between public concern and policy response is widening. Experts warn new models pose critical security risks. The public is increasingly skeptical of AI hype. Yet the policy direction is full acceleration, no brakes. This isn't about Trump specifically. It's about a regulatory environment where the only argument that matters is "but China."
The Implication
Big tech just learned it can win any AI policy fight by making it about China. Expect that frame to dominate every future debate about model safety, compute governance, or agent deployment standards. If you're building in this space, the message is clear: ship fast, apologize later, invoke geopolitical competition if anyone asks questions.
For everyone else, this is the shape of Web4 infrastructure being set in real time. The agent economy will run on models that face zero mandatory safety gates before deployment. That's not a prediction, it's policy now. Plan accordingly.