AI companies look to the ocean as a place to put more data centers

AI companies are increasingly looking to the ocean as a potential site for new data centers, driven by the need to address challenges related to energy consumption, water use, and carbon emissions associated with traditional land-based data centers. Microsoft, for instance, conducted a two-year research project exploring the feasibility of underwater data centers, concluding in 2024 without plans to build more.

The project, however, showed promising results, with underwater servers experiencing failure rates about one-eighth of those in similar land-based centers.

China recently launched what may be the world's first wind-powered underwater data center in Shanghai, which uses seawater for cooling and claims to use at least 30% less electricity than traditional data centers. Japan is also testing the use of floating data centers in containers on the sea, which could provide unobstructed conditions for solar panels and wind turbines and reduce land demands. Singapore is developing a similar floating data center to avoid using limited land.

South Korea is planning an underwater data center in Ulsan, which would use 30% less power than land-based centers by utilizing seawater for cooling. Meanwhile, in Maine, a submersible AI data center is proposed to be powered by tidal turbines in the Bay of Fundy. These various approaches aim to reduce demand for grid-supplied electricity and water, as well as potentially lower users' distance from data centers.

However, challenges remain, particularly regarding maintenance and the environmental impacts of discharging warm water into the ocean.