The Zhengzhou cluster does not merely demonstrate that China can build large AI systems without US chips — it establishes that the assumption underwriting US export policy is now empirically false. The policy logic held that denying access to leading-edge accelerators would throttle Chinese AI development at the compute layer. That logic required Chinese domestic alternatives to remain meaningfully below the threshold of frontier-scale deployment. The 60,000-chip activation in Zhengzhou and the parallel Shenzhen deployments collectively demonstrate that the threshold has been crossed without US hardware. The institutions that shaped…