Panthalassa, a Portland-area startup that wants to run artificial intelligence on wave-powered floating platforms, has hauled in a $140 million Series B to move from prototypes to production and deploy its Ocean‑3 nodes. The company says the new cash will finish a pilot factory near Portland and speed up at-sea pilots that use wave energy to power AI inference chips, with results beamed back to shore by satellite. The raise puts a local spotlight on one of the more ambitious attempts to ease the national crunch on electricity and cooling that is squeezing land-based AI data centers.
In a press release via Business Wire, Panthalassa said the round was led by Peter Thiel, with participation from John Doerr, TIME Ventures, SciFi Ventures, Hanwha Group, Fortescue Ventures and a long roster of other strategic and venture investors. The company says the capital will complete its pilot manufacturing facility near Portland and accelerate the Ocean‑3 pilot node series in the northern Pacific this year, setting up for commercial deployments in 2027. “We’re now ready to build factories, deploy fleets, and provide a sustainable new source of energy for humanity,” CEO Garth Sheldon‑Coulson said in the release.
How The Ocean‑3 Nodes Run AI Offshore
Each Ocean‑3 node is described as an autonomous plate‑steel float that rides incoming waves, forcing seawater through an internal turbine to generate electricity. Instead of sending that power back to shore, the node uses it on board to run AI inference hardware. The surrounding ocean acts as passive super‑cooling for the chips, and only the inference results are sent to land over satellite links. Independent reporting notes that the design could deliver relatively cheap, continuous power, but also faces classic marine headaches like corrosion, biofouling and storm damage, as reported by CBS News.
Local Factory Push And Timeline
Panthalassa says the Series B will bring its pilot manufacturing facility near Portland across the finish line, with an eye toward mass‑producing nodes in coastal factories and pushing the design toward industrial scale. The company points to prior sea tests of its Ocean‑1, Ocean‑2 and Wavehopper prototypes in 2021 and 2024, and it plans to stage Ocean‑3 pilots in the northern Pacific in 2026 before targeting commercial roll‑outs in 2027, according to Investing.com.
Investors And The Pitch
Backers frame Panthalassa as a way to sidestep land-based bottlenecks such as grid capacity, water for cooling and lengthy permitting by putting compute where the energy already is. In the company release via Business Wire, Peter Thiel said, “The future demands more compute than we can imagine,” while investor John Doerr called the plan “a triple win” for workers, communities and U.S. technological leadership.
Questions And Risks
Reporters and engineers are quick to point out that the ocean does not make anything easy. Salt, storms and long repair cycles can turn a promising prototype into a maintenance nightmare, and satellite links introduce latency that limits which AI workloads make sense offshore. If a node fails far out at sea, retrieval or repair could cost significantly more than fixes at a land facility, and the business case will hinge on durability and reliable offtake agreements. Those tradeoffs surface in coverage that weighs the technology’s upside against practical challenges of scaling, according to DataCenterDynamics.
What To Watch
Key milestones to track include completion of the Portland-area pilot factory, the first Ocean‑3 pilot runs in the northern Pacific and any commercial offtake deals with cloud or AI companies that would validate the model. If the factories land orders and the pilot nodes survive rough seasons at sea, Panthalassa’s concept shifts from eye‑catching idea toward industrial supply chain, a progression investors and local manufacturers will be watching closely, analysts say via TipRanks.
