Startup Drogue Raises $20 Million to Power Floating Data Centers With Ocean Heat

Drogue, a Washington, D.C. startup founded just six months ago, has raised $20 million to build floating data centers powered by heat drawn from the ocean, with a multi-megawatt prototype scheduled to go to sea in early 2028.

The company told TechCrunch exclusively that the round includes XYZ Venture Capital, Toyota Ventures, Tamarack Global, Dream, FDVC, Garuda Ventures, Flexcap, Boost VC and Slow Ventures, among other investors. Drogue has stayed quiet until now. It has already secured "approval in principle" — a preliminary sign-off — from the American Bureau of Shipping, the body that sets and checks safety standards for ships, and it is now buying its first ship to start building the open-ocean prototype.

Why put data centers at sea?

Co-founder and CEO Sam Sipe, a former U.S. Navy aviator who flew E-2D Hawkeyes off aircraft carriers, frames the problem as a mismatch of clocks. "The data center industry is a real estate game, and so they're playing on 8- to 10-year timelines, including interoperability with utilities and [independent system operators]," he said. "AI demand has increased 10x in just the last three years. It's completely incompatible."

Drogue's answer is to place servers on vessels that carry their own power source. Each vessel would stay at sea for about a decade. The first prototype will ride on a traditional oceangoing ship; later versions would sit on floating structures that resemble offshore oil platforms.

Data centers suit the marine setting better than most industries, according to Sipe. Sending power ashore would need long, costly cables, and on-board manufacturing would require expensive shipping of materials and finished goods. Servers, by contrast, are compact, and moving data off the platform costs almost nothing. In plain terms: the ocean holds the power, and data is the cheapest cargo to bring back.

How does the system work?

The technology draws on ocean thermal energy conversion, or OTEC, a method first developed in the 1950s and tested in the 1980s before being set aside. It works, in effect, like a heat pump run in reverse. A heat pump uses electricity to move heat from one place to another; Drogue's system uses a temperature difference to generate electricity.

The mechanism runs on the gap between sun-warmed surface water and the deep ocean, which averages about 39°F (4°C). The system:

• draws in warm surface water and uses it to cool the data centers;
• sends the now-warmer water to boil ammonia, a chemical that turns to gas at -28°F;
• passes the ammonia gas through a turbine to generate electricity;
• runs the gas through a condenser, where cold water pumped from 1,000 meters deep absorbs the heat;
• returns the warmed water to 1,000 meters, where it dissipates, and recycles the re-liquefied ammonia through the sealed loop.

"There's basically no amount of heat we can pull out of that that will collapse that temperature differential," Sipe said. The scale backs him up: as the planet warms, the oceans absorb an additional 5.8 million terawatt-hours of energy every year, against roughly 180,000 terawatt-hours consumed annually by all of human civilization.

Why now, and why did it fail before?

The idea was tested in the 1980s and then dropped. Sipe tracked down engineers from those early efforts and interviewed them about what went wrong. "And everyone was like, dude, this stuff works. It's just a matter of being able to raise enough capital to build something big enough, because there's a sweet spot," he said.

He points to advances since the 1980s, specifically in manufacturing techniques for the pipe and the heat exchangers, as the reason the old economics no longer apply. His view of the ocean itself is similarly contrarian. "Most people think, oh, ocean — hard waves, storms, corrosion, biofouling. And I just see this majestic place and resource that we just ignore," he said.

How does it compare with orbital data centers?

The concept mirrors Elon Musk's plan for orbital data centers in one respect: find cheap power and put the data center next to it. The difference is transport. Orbital facilities need expensive rocket launches. Drogue's vessels simply sail out to sea, something ships have done for millennia.

The first prototype, a multi-megawatt unit, should be in the water and crunching data in less than two years, according to Sipe. Whether OTEC can be scaled from a 1980s curiosity into a commercial power source for AI workloads will be answered when the first ship sails in early 2028.