Can AI go underwater?

Underwater data centers represent a strategic response to the massive energy and water demands of the global AI boom. By leveraging natural seawater for cooling and integrating with offshore wind farms, these facilities drastically reduce electricity consumption and eliminate the need for freshwater cooling, which typically consumes thousands of tons annually. While projects like the Shanghai-based underwater data center demonstrate significant land-use efficiency and carbon reduction, they face substantial technical challenges, including high construction costs, complex maintenance requirements, and the need for advanced remote monitoring. Beyond underwater solutions, the industry is exploring space-based computing and utilizing naturally cold climates to mitigate the environmental impact of traditional land-based data centers. These efforts reflect a broader shift toward balancing rapid digital expansion with sustainable infrastructure as the global reliance on high-performance computing continues to grow.