China Ai Data Centre: Servers in the ocean: World’s first offshore underwater AI data centre launched in China | World News

China has officially launched the world’s first commercial offshore wind-powered underwater AI data centre off the coast of Shanghai’s Lingang Special Area. The futuristic facility contains nearly 2,000 servers housed inside pressure-resistant underwater modules positioned beside offshore wind turbines. Designed to support artificial intelligence workloads, big data processing and domestic large language model development, the project combines renewable energy with natural seawater cooling to reduce electricity use, land requirements and freshwater consumption. Chinese developers say the underwater system achieves a highly efficient Power Usage Effectiveness (PUE) rating of around 1.15, significantly lower than many traditional land-based data centres.

A massive underwater computing project near Shanghai

The project was launched through a partnership involving the Lingang Special Area Administrative Committee, Shanghai Lingang Special Area Investment Holding Group and HiCloud Technology. Additional operational agreements were later signed with companies including China Telecom, Shenergy Group and CCCC Third Harbor Engineering.Construction on the facility began after agreements were finalised in June 2025. The project reached construction completion in October 2025 before beginning trial operations earlier this year. Full commercial operations officially started in May 2026.Located between the first and second phases of Lingang’s offshore wind farm, the data centre directly connects to nearby renewable energy infrastructure. The entire project reportedly cost around ¥1.6 billion, or approximately $226 million.

Nearly 2,000 servers deployed underwater

The underwater data centre was developed in phases, beginning with a smaller 2.3 MW demonstration facility before scaling up to a full 24 MW commercial operation.Inside the subsea modules are nearly 2,000 servers designed to support artificial intelligence computing, big data annotation, 5G infrastructure services, domestic large language model training and cloud computing operations. China Telecom and local computing provider LinkWise have already deployed GPU clusters inside the underwater modules to support AI processing and data-intensive workloads.Unlike conventional server farms built inside massive industrial buildings, the Lingang facility uses sealed underwater capsules positioned beneath the ocean surface close to offshore wind turbines. Developers say this design helps reduce land usage while improving cooling efficiency.

How seawater cooling works

One of the project’s most important innovations is its cooling system. Traditional data centres rely heavily on industrial air-conditioning systems and cooling towers to prevent servers from overheating. Those systems consume enormous amounts of electricity and often require substantial freshwater supplies.The Shanghai underwater data centre instead uses seawater as a passive cooling mechanism. According to HiCloud Technology, server heat changes refrigerant inside copper pipes from liquid into gas. The gas naturally rises to an upper cooling layer, where surrounding seawater removes heat through a heat exchanger. Once cooled, the refrigerant condenses back into liquid form and returns to the server modules through gravity.Developers say this creates a low-energy heat exchange cycle that significantly reduces the need for conventional powered cooling systems.

Why the PUE rating matters

The facility reportedly maintains a Power Usage Effectiveness rating of around 1.15.PUE measures how efficiently a data centre uses electricity. A lower score means more energy is going directly toward computing rather than cooling or infrastructure overhead.Many traditional enterprise data centres operate closer to 1.5 or even 2.0 because of heavy cooling demands. By comparison, a PUE of 1.15 places the Lingang underwater facility among the most energy-efficient large-scale data centres currently operating.Project developers claim the system reduces electricity consumption by 22.8%, eliminates freshwater use entirely and cuts land usage requirements by more than 90%. These figures are especially significant at a time when AI infrastructure is rapidly increasing global electricity and water consumption.

Powered by offshore wind energy

Another major feature of the project is its renewable energy integration. The underwater modules are positioned next to Lingang’s offshore wind farms, allowing the facility to draw most of its electricity directly from wind energy generated at sea.Reports suggest more than 95% of the project’s electricity comes from renewable energy sources. This is particularly important as artificial intelligence systems and GPU clusters continue to increase global power demand at a rapid pace.China’s underwater project combines renewable energy, subsea engineering, artificial intelligence computing and natural cooling systems into a single integrated platform. The project is widely viewed as part of China’s broader push to strengthen its position in AI infrastructure and green computing technologies.

The engineering challenges underwater

Despite its efficiency advantages, underwater infrastructure introduces several engineering difficulties. Saltwater corrosion remains a major long-term concern because seawater can gradually damage electronics and metal structures. Engineers must also ensure the underwater capsules remain completely sealed against pressure and water intrusion for years at a time.Maintenance is another major challenge. Replacing damaged hardware underwater is far more complicated than servicing traditional server racks inside ordinary buildings. Operators therefore rely heavily on sealed modular systems, remote monitoring and infrastructure redundancy to minimise the need for physical repairs.Subsea fibre-optic and power cables also present reliability concerns because underwater cable repairs are both expensive and technically difficult.

Comparisons with Microsoft’s Project Natick

The Shanghai project has drawn comparisons with Microsoft’s experimental Project Natick programme, which tested underwater data centre capsules off the coast of California in 2015 and later near Scotland’s Orkney Islands in 2018.Microsoft’s experiments showed underwater environments could reduce hardware failure rates because stable temperatures and sealed conditions limited exposure to oxygen and human interference. However, Microsoft ended the project commercially by 2024 without moving to large-scale deployment, partly because maintenance and operational complexity remained difficult.China’s Lingang facility differs because it has already entered full commercial operation at a significantly larger scale and is directly integrated with offshore renewable energy infrastructure.Although the Lingang project is widely described as the world’s first offshore wind-powered underwater data centre, it is not the first underwater data centre overall.That distinction belongs to an earlier commercial underwater data centre project launched near Hainan, China. What makes the Shanghai facility unique is its integration of offshore wind power with large-scale underwater AI computing infrastructure.

Why underwater data centres could become more common

The rapid growth of artificial intelligence is dramatically increasing global demand for electricity, cooling systems and land for data centres. Technology companies and governments are now exploring alternatives to conventional server infrastructure, including underwater data centres, floating computing platforms, liquid immersion cooling systems and even nuclear-powered AI facilities.China’s underwater project represents one of the most ambitious attempts so far to combine renewable energy generation with sustainable artificial intelligence computing infrastructure. If the Lingang system performs successfully over the coming years, similar offshore underwater facilities could become an important part of future global AI infrastructure.