China has begun full commercial operation of an underwater data center located more than 30 feet beneath the East China Sea off Shanghai’s Lingang Special Area. The $226 million facility represents a shift in how operators cool and power large-scale computing infrastructure in densely populated regions.
The underwater data center operates between two phases of an offshore wind farm, drawing 95 percent of its electricity from wind generation. The facility uses seawater for passive cooling instead of traditional industrial chillers, a design choice that significantly reduces energy consumption.
Facility Specifications and Capacity
The 24 megawatt underwater data center houses nearly 2,000 servers, including GPU clusters from China Telecom and LinkWise. The facility is designed to handle artificial intelligence workloads, reflecting growing demand for computing power in machine learning applications.
According to developers, the system cuts electricity consumption by 22.8 percent compared to conventional facilities. The design eliminates freshwater use entirely and reduces land use by more than 90 percent, addressing resource constraints in coastal industrial areas.
Underwater Data Center Performance Metrics
The facility operates at a PUE (Power Usage Effectiveness) of 1.15, a metric that measures how much total energy a data center consumes relative to the energy delivered to computing equipment. Lower PUE values indicate greater efficiency. Microsoft’s Project Natick previously demonstrated that submerged servers can be up to 8 times more reliable than land-based servers, though Microsoft discontinued that research program.
Passive seawater cooling eliminates the need for mechanical refrigeration systems, reducing both operational costs and carbon emissions. This approach addresses a major expense in data center operations across warm climates.
Technical Challenges in Underwater Operations
The primary challenge with underwater data center deployment involves equipment maintenance. When server components fail at depths exceeding 30 feet, repairs become logistically complex and costly. Operators must use specialized diving equipment and procedures to access and replace failed hardware, extending downtime and repair expenses.
This maintenance constraint was a key factor in Microsoft’s decision to halt Project Natick, despite the reliability advantages demonstrated by submerged hardware. Balancing operational efficiency gains against maintenance accessibility remains a central engineering consideration for underwater deployments.
Regional and Global Implications
The Shanghai facility demonstrates how offshore infrastructure can integrate renewable energy sources with data security and cooling requirements. As demand for data center capacity continues to grow, particularly for AI applications, operators are exploring non-traditional locations and designs to manage environmental impact and power costs.
China’s investment in underwater data center technology reflects broader efforts to support high-performance computing infrastructure while addressing constraints in land availability and freshwater resources in coastal regions. The success or challenges of this facility will likely influence how other operators evaluate similar deployments in comparable geographic and climate conditions.
