China's AI push is aimed at economic growth rather than a race to build a superintelligent machine, which undercuts the main argument US companies use to resist any rules — and makes a deal with Beijing more possible than Washington admits.

• Chinese leaders mostly see AI as an engine for growth and productivity, not a sprint to artificial general intelligence like Silicon Valley's.
• Chinese models are now close enough to the best US models that the gap is almost gone, driven by cheap, open models anyone can download and build on.
• Beijing cares intensely about near-term risks — social stability, critical infrastructure, cyber attacks — and has shown it will rein in its own tech companies when it wants to.
• The Trump-Xi meeting produced real if modest wins: continued AI safety talks and a hotline for reporting serious AI incidents.
• The "we must build it before China does" pitch from Anthropic's Dario Amodei and OpenAI's Sam Altman reads as a business argument dressed up as safety, making good-faith talks harder.

Outlook: Expect small, incremental US-China AI agreements to build trust rather than any sweeping pact, while Chinese open models keep closing the capability gap.