Plot twist: 'locked-down' China just out-opened America on AI
What happened
A new op-ed argues China's strategy of releasing open-weights AI models is quietly winning the global AI race, while US labs like OpenAI and Anthropic stay locked down behind APIs and enterprise contracts. The piece cites a16z's Martin Casado, who claims there's an 80% chance any given startup is using a Chinese model somewhere in its stack.
Why this matters
If the moat isn't the model but the ecosystem around it, then whoever gets their weights into the most hands wins long-term — and that's increasingly China, not the US. Export controls on GPUs pushed Chinese labs toward openness as a distribution strategy, accidentally out-maneuvering America's closed-shop playbook.
The slightly cynical read
Nothing says "national AI strategy" like the US restricting chips to China and China responding by just... giving away the software for free. Turns out you can't export-control your way out of an open-source flywheel.
What to watch next
Watch whether OpenAI, Anthropic, or Meta respond with more aggressive open-weight releases to compete on distribution, not just benchmarks. If Casado's 80% figure holds up under scrutiny, expect a lot of uncomfortable congressional hearings about "AI sovereignty."