AI data centers may soon ride ocean waves

Silicon Valley investors are backing Panthalassa, a company aiming to deploy autonomous, floating AI computing nodes at sea. The company recently secured $140 million in Series B funding, led by Peter Thiel, bringing its total funding to $210 million. This investment will support the completion of a pilot manufacturing facility near Portland, Oregon, with the first Ocean-3 pilot node series slated for deployment in the northern Pacific Ocean by 2026. Panthalassa's innovative approach seeks to mitigate the challenges of land-based AI data centers, which require vast amounts of electricity, space, and cooling, often facing community pushback. Panthalassa's design leverages ocean waves for electricity generation and seawater for cooling its onboard computing chips, which are intended for AI inference tasks. The company has spent a decade developing its wave-energy conversion and autonomous ocean operation technology, testing prototypes like Ocean-1, Ocean-2, and Wavehopper. CEO Garth Sheldon-Coulson highlights the platform's ability to operate in energy-dense wave regions far from shore, converting that resource into reliable clean power. The company believes this method could reduce pressure on terrestrial energy grids and water resources. However, the concept faces practical challenges, including reliance on potentially slower low-Earth-orbit satellite links for data transmission, difficulties in coordinating multiple nodes, and the complexities of maintenance in harsh ocean environments. Panthalassa plans to demonstrate AI inference capabilities and refine manufacturing before commercial deployments in 2027. While previous efforts like Microsoft's Project Natick explored underwater data centers, Panthalassa's unique combination of wave power, onboard AI, and satellite communication distinguishes its approach, though skepticism remains regarding its long-term viability and the potential for new environmental and governance questions in international waters.