Shanghai’s Commission of Economy and Informatization issues a software and information‑services plan to create a tiered supply system of large clusters + small clusters + edge compute. It calls for 100,000‑GPU‑class ultra‑large AI compute clusters in Songjiang, Lingang and Qingpu, and thousand‑GPU clusters in Baoshan, Pudong and Jiading. The plan directs conversion of traditional data centers and telecom rooms into ~100‑GPU edge compute centers to meet enterprise and consumer ultra‑low‑latency c

Shanghai’s Commission of Economy and Informatization issues a software and information‑services plan to create a tiered supply system of large clusters + small clusters + edge compute. It calls for 100,000‑GPU‑class ultra‑large AI compute clusters in Songjiang, Lingang and Qingpu, and thousand‑GPU clusters in Baoshan, Pudong and Jiading. The plan directs conversion of traditional data centers and telecom rooms into ~100‑GPU edge compute centers to meet enterprise and consumer ultra‑low‑latency compute needs. It prioritizes finance, education, healthcare, culture/tourism and industry for Model‑as‑a‑Service (MaaS) platforms offering industry app marketplaces, model customization and hosting, Agent construction, low‑code development, APIs, compute provisioning and management, and AI inference services.