The world's two AI superpowers just spent three days proving they care more about not losing than about not losing control.
The Summary
- Trump and Xi wrapped their Washington summit with zero progress on AI governance, choosing pageantry over policy while their respective AI programs accelerate unchecked
- Trump signaled on social media he would avoid any agreement limiting AI development, even as Congress debates a bill to ban artificial superintelligence
- The only concrete outcome: a modest two-month trade truce extension, while both nations' AI arms race continues at full throttle
The Signal
The three-day state visit was heavy on ceremony and light on substance. Xi toured the National Archives, viewed the Declaration of Independence, and exchanged pleasantries with Trump about Iran, Ukraine, and Taiwan. What they didn't do: establish any framework for coordinating on AI safety, share research on alignment, or even acknowledge that their race to artificial general intelligence might benefit from guardrails.
Trump telegraphed his position before the talks even started, posting that AI would be a "major part" of discussions but suggesting he'd reject any bilateral limits on development. The message was clear: America builds, China builds, nobody blinks first.
"The US president emphasized personal rapport with Xi rather than substantial agreements to pause their countries' AI arms race."
Meanwhile, back in Washington, Congress is considering legislation to ban artificial superintelligence entirely. The timing is almost comedic. The executive branch refuses to coordinate with Beijing on AI development while the legislative branch floats the idea of banning the most advanced versions outright. Congress has done little to regulate AI despite warnings from industry leaders about catastrophic risks.
The dynamic exposes the core coordination problem of the agent economy:
- Both nations believe slowing down means losing
- Both nations know unchecked development carries existential risk
- Neither nation trusts the other enough to verify compliance with any agreement
Xi reportedly wanted Trump to halt arms sales to Taiwan. Trump wanted Xi to pressure Iran. What neither wanted, apparently, was to be the first to suggest their AI labs might benefit from adult supervision. The result is a summit that changed nothing about the trajectory both countries are on.
The Implication
Watch what happens when the Congressional ASI ban gets serious attention. If it gains momentum, expect the White House to frame any restrictions as unilateral disarmament. The argument will be: China won't limit their research, so neither can we. It's the same logic that drove Cold War weapons stockpiling, now applied to training runs and parameter counts.
For companies building in the agent economy, this non-agreement is actually clarity. No international framework is coming to slow things down. The regulatory environment will remain fragmented, nationalist, and driven by competitive fear rather than coordinated caution. Build accordingly.