Overview
- The Shanghai Lingang undersea data centre began commercial operation in May and sits about 10 metres below the sea surface off the Lin‑gang area with a 24 MW capacity and roughly 2,000 servers.
- Project developers say the facility draws more than 95% of its electricity from a nearby offshore wind farm and achieves a power usage effectiveness near 1.15, cutting electrical demand by about 20–23% versus typical land sites.
- China Telecom has installed GPU clusters in the centre to run AI workloads, marking the site as an operational testbed for high‑performance computing at sea.
- Engineers and independent scientists warn of unresolved risks including long‑term saltwater corrosion, subsea cable reliability, limited access for hardware replacement, seabed disturbance, localized warming and possible effects such as algal blooms.
- The project cost about ¥1.6 billion (≈US$225–226 million) and is presented as a proof‑of‑concept rather than a large‑scale solution for AI demand, building on earlier pilots like Microsoft’s Project Natick while highlighting regulatory and scaling hurdles for wider adoption.