China is advancing its artificial intelligence ambitions by turning to an unconventional solution: underwater data centres. The country is actively developing and deploying submerged computing facilities to support the growing demand for AI processing power while addressing the rising energy and environmental challenges associated with traditional data infrastructure.
The initiative involves placing sealed server units beneath the ocean surface, where natural seawater is used to cool high-performance computing systems. This approach significantly reduces the need for energy-intensive cooling mechanisms that conventional land-based data centres rely on. Cooling alone can account for nearly 40 percent of a data centre’s electricity consumption, making efficiency improvements critical as AI workloads expand.
One of the most prominent projects is being developed off the coast of Shanghai, where construction began on a wind-powered underwater data centre located about six miles offshore. The facility is designed to leverage both seawater cooling and renewable energy sources such as offshore wind, reflecting China’s broader push toward sustainable digital infrastructure.
China has already made significant progress in this space, including a $226 million underwater data centre project in Shanghai’s Lin-gang Special Area. The facility uses ocean currents for cooling and is expected to drastically cut energy consumption compared to traditional data centres. Some estimates suggest that underwater setups could reduce cooling-related energy use by up to 90 percent, highlighting their potential efficiency advantage.
Beyond sustainability, these underwater facilities are being designed to support large-scale AI workloads. For instance, an underwater data centre in Hainan has been expanded with additional modules capable of processing up to 7,000 AI queries per second, demonstrating the growing computational capacity of such systems.
The shift toward underwater infrastructure also reflects China’s strategy to overcome land and resource constraints while scaling its AI ecosystem. By reducing reliance on freshwater and land space, these projects aim to provide a more sustainable path for expanding computing capacity in the face of surging global demand for AI-driven services.
However, the technology is still evolving and faces challenges, including maintenance complexity, environmental concerns, and long-term scalability. Despite these hurdles, China’s rapid progress in underwater data centres signals a bold rethinking of how future AI infrastructure can be built, potentially influencing global approaches to sustainable computing.
