As of October 2026, Zhejiang province has built 113 computing centers with a deployed total compute capacity exceeding 198 EFLOPS, 96 percent of which is intelligent computing power, Caixin-linked financial outlet CLS reported on November 14. Five supercomputing clusters of at least 10,000 GPU-scale cards, including the Pinghu Runze facility, are already in service, and field reporting in the same story found AI applications reaching villages, hospitals and traffic intersections across the province.
The numbers land against a clear policy backdrop. China's 15th Five-Year Plan outline calls for accelerating a unified national computing network and fully implementing the "AI Plus" initiative, formally placing computing infrastructure inside the country's core matrix of new infrastructure. For Zhejiang, an early mover in the digital economy, the newly disclosed figures amount to an interim balance sheet of that build-out.
According to the report, Zhejiang's next steps focus on three fronts: deepening AI integration with manufacturing, expanding application scenarios and the industrial ecosystem, and completing a support system that coordinates computing power, data and models. The emphasis falls on lowering the threshold and cost of using AI for enterprises, especially smaller ones.
The regional race to publish bigger compute numbers is entering a second stage. After 113 centers and a 96 percent intelligent-computing share, the deciding variables are scheduling efficiency and the electricity-and-services bill: whether small and mid-sized firms can actually afford to switch AI on is the real test of whether this capacity gets used.