Geographic Freedom Enables Offshore Compute
Panthalassa
Panthalassa’s real edge is not just cheaper power, it is the ability to put compute where physics works best instead of where grid politics allow. Land data centers have to secure substation capacity, water, permits, and neighbor approval before a single server turns on. Panthalassa can instead chase high wave energy in the northern Pacific, use seawater for cooling by default, and pair deployment with coastal steel fabrication near Portland, which makes expansion look more like shipbuilding than real estate development.
-
The manufacturing plan matters as much as the ocean plan. The Series B funds a pilot factory near Portland, and the product is described as mass produced plate steel nodes. If one factory can repeatedly turn out working units, growth can come from cloning factories along coastlines, then towing nodes into nearby deployment corridors.
-
Fluidstack shows what buyers reward in this market. It moved from GPU rental into acquiring power, building data centers, and operating them, then won a $50B Anthropic infrastructure deal. That is the template Panthalassa is aiming at, customers sign long duration capacity contracts when uptime and delivery are credible, not when energy is merely novel.
-
Supermicro’s involvement points to a second bottleneck beyond energy, server packaging. Offshore compute needs racks, cooling, and enclosures that survive salt, motion, and remote operation. Supermicro already sells pre validated AI systems at rack and data center scale, so a partnership could shorten the path from ocean platform to usable compute product.
If Ocean-3 proves stable uptime, Panthalassa can turn geographic freedom into a rollout machine, factory by factory and corridor by corridor. The next phase is a shift from proving wave power works to proving offshore compute can be contracted, manufactured, and serviced with the predictability buyers already expect from large terrestrial AI infrastructure providers.