The pitch sounds almost absurd: Drop massive, lollipop-shaped computing platforms into the open ocean, let them harvest power from the waves, cool their processors in seawater, and beam results back to land via satellite. No power lines. No fuel shipments. Just the ceaseless heave of the Pacific doing the work that landlocked data centers can no longer handle.
Yet on May 4, 2026, Peter Thiel and a constellation of heavyweight investors decided the idea was worth $140 million.
Panthalassa, the Portland startup behind the concept, announced its Series B that morning with Thiel's backing pushing the company to a valuation hovering around $1 billion, according to Axios Pro Rata. It's a remarkable sum for technology that has yet to prove it can survive the ocean's fury at commercial scale—but it's also a measure of just how desperate the AI industry has become for power.
The Calculus Behind the Check
Thiel didn't come alone. The syndicate reads like a climate-tech fever dream: John Doerr, the venture legend who backed Google and Amazon; Marc Benioff's TIME Ventures; Max Levchin's SciFi Ventures; and Hanwha Group, the South Korean conglomerate. Australian mining billionaire Anthony Pratt joined. So did Fortescue Ventures and Super Micro Computer, the server maker whose hardware presumably would end up bobbing in those distant swells.
Returning investors—Founders Fund, Gigascale Capital, Lowercarbon Capital, Unless, and WovenEarth—doubled their bets. In all, more than 20 backers participated, from Portland Seed Fund to Tokyo's Sozo Ventures and Susquehanna Sustainable Investments.
Press coverage states approximately $210 million invested so far as of May 5–6, 2026, though the company has declined to confirm the total.
The math driving the round is stark. Morgan Stanley Research published in March–April 2026 projects U.S. data center power demand will reach roughly 74 gigawatts by 2028, with an estimated shortfall of 49 gigawatts—a gap fueling over $1 trillion in infrastructure commitments. Solar farms take years to permit. Nuclear plants take longer. The open ocean, Panthalassa argues, is available now.
"Solar, nuclear, and the open ocean" are the only sources with "tens of terawatts of new capacity potential," CEO Garth Sheldon-Coulson said in the funding announcement. It's a sweeping claim for a technology still in sea trials, but perhaps not more sweeping than the problem it aims to solve.
What Gets Deployed in 2026
The money will finish a pilot manufacturing facility near Portland, accelerate deployment of Ocean-3 prototype nodes in the northern Pacific in 2026, and bankroll commercial rollouts targeted for 2027. That timeline is aggressive. Microsoft famously tested an underwater data center concept—Project Natick—only to shutter it after mixed results. China has experimented with subsea computing off Shanghai. The ocean, it turns out, is unforgiving to complex machinery.
Panthalassa's Ocean-3 units are engineered around that harshness. Each device stretches roughly 85 meters in length, according to Financial Times reporting cited by Ars Technica—a spherical "head" atop a long submerged tube. Wave motion drives internal turbines that generate electricity. AI chips mounted in the floating platform use that power for inference tasks, while the surrounding seawater offers what the company calls "free supercooling."
Results travel back via low-Earth-orbit satellites, a design choice that introduces latency tradeoffs. Real-time applications are likely out; batch processing is the sweet spot. The nodes are autonomous, self-propelled, and built to be stamped out from plate steel in coastal shipyards—industrial manufacturing at scale, not boutique engineering.
Sheldon-Coulson, a former Bridgewater Associates analyst who holds multiple ocean energy patents, founded Panthalassa as a public benefit corporation in 2016. The company has been testing prototypes since 2021; its Ocean-2 unit completed a three-week trial off Washington state in February 2024.
The Frontier Logic
Thiel, who has backed everything from seasteading to anti-aging startups, framed Panthalassa as opening "the ocean frontier." Doerr called the wave-power system a "game changer." Both sound enthusiastic, though neither has explained what happens if a typhoon tears through a deployment zone or if maintenance costs spiral beyond projections.
The company has not disclosed commercial customers or letters of intent. The pilot nodes scheduled for this year will test whether the technology can withstand months at sea under real conditions—corrosion, biofouling, storms, and the mechanical strain of constant motion.
If Panthalassa succeeds, it will have cracked a physics problem that unlocks energy where power grids don't reach. If it stumbles, $210 million will have bought an expensive lesson in the gap between venture-scale ambition and ocean-scale reality. The 2027 commercial deadline leaves little margin for error.
For now, the bet is placed. The Pacific will render its verdict soon enough.
