Startup Raises $140 Million to Deploy Wave-Powered Floating AI Data Centers

A company developing floating AI data centers has raised $140 million in Series B funding to advance its plan to place computing power at sea. The round brings the company's total funding to $210 million and was led by an investor focused on frontier technologies.

The money will be used to complete a pilot manufacturing facility near Portland, Oregon, and to deploy the first Ocean-3 pilot nodes in the northern Pacific Ocean later this year. The floating nodes are designed to capture wave motion to generate electricity while using surrounding seawater to cool onboard servers.

The company has spent a decade refining the technology for power generation, computing and autonomous operations at sea. Earlier prototypes, Ocean-1 and Ocean-2, were tested in 2021 and 2024. Instead of relying on land-based power grids, the nodes would produce their own electricity from waves in energy-dense regions far from shore.

The systems would perform AI inference tasks, responding to prompts after models have been trained elsewhere. Results would be sent back to users via low-Earth-orbit satellite links.

Each node consists of a surface structure that bobs with the waves and a long submerged tube that oscillates to drive internal turbines. The AI servers are housed in a sealed compartment cooled by seawater. This approach aims to address the massive electricity and cooling demands of AI data centers without drawing on terrestrial infrastructure.

The company plans commercial deployment in 2027 after demonstrating reliable AI inference with the Ocean-3 series. However, operating far offshore introduces limitations. Satellite connections provide lower bandwidth and higher latency than fiber-optic cables used by land-based data centers, making coordination between multiple nodes more difficult.

Repair and maintenance present additional hurdles. While land facilities can dispatch technicians quickly, floating nodes in rough seas would require specialized vessels and favorable weather windows. Saltwater corrosion, storms and constant motion add stress to hardware and could increase long-term operational costs.

Land-based facilities face constraints including limited grid capacity, community opposition and high cooling costs. The company argues its approach could ease pressure on electrical grids and reduce the need for new power plants onshore. Previous offshore experiments have shown mixed results.

One major technology company tested submerged servers using seawater for cooling and reported lower failure rates than land-based counterparts before ending the project. Other efforts in Asia have explored similar underwater concepts near coastal cities where land is scarce.

The latest funding reflects increased investor interest in unconventional compute infrastructure as demand for AI continues to surge. The company states it is now ready to build factories and deploy fleets, positioning ocean wave energy alongside solar and nuclear as a major new source of clean power.

>"The future demands more compute than we can imagine. Extra-terrestrial solutions are no longer science fiction. Any impact would emerge gradually if the nodes prove reliable at scale and help meet rising electricity demand without straining local infrastructure.

Success could influence debates over marine environmental oversight and governance of computing resources in international waters. The company must still demonstrate that its autonomous systems can operate for years in harsh ocean conditions with minimal human intervention.

If successful, the approach could expand available capacity for certain AI workloads. If not, it would underscore the practical difficulties of moving critical infrastructure offshore.