Garth Sheldon-Coulson

Co-Founder & CEO, Panthalassa · 1 appearance on the record.

computed by AI from the episodes · how this works → · full disclaimer →

founderexecutivescientistanalyst@garthsc ↗LinkedIn ↗panthalassa.com ↗

Garth Sheldon-Coulson is the co-founder and CEO of Panthalassa, a startup building wave-powered floating nodes for offshore AI inference data centers cooled by the deep ocean. Before co-founding Panthalassa in 2016, he worked as an AI and energy researcher and analyst at Bridgewater Associates.

18statements → 12claims → 4claims resolved → 75%fully supported → 3.78/5average certainty → 2.06/5average debate potential → ≈4.5/5argument clarity, estimated →

3 supported 1 partly supported 0 contradicted 1 not yet assessed 7 not checkable as stated how the 12 claims stand · each chip opens the sources

3 predictions · 9 assertions · 1 insight · 5 disclosures · every statement was checked. The predictions and assertions are the 12 claims: statements the public record can support or contradict. 4 are resolved, 1 is not yet assessed, and 7 name no date, number or outcome precise enough to check. Everything else (opinions, insights, what ifs, disclosures) can never be settled by the record, so it carries no assessment.

The record, in short

What the tape says about how Garth argues and how the claims held up. Everything they said, and everything said about them, is in the tabs below.

Their most notable supported claim

Assertion Supported
Panthalassa nodes passively self-propel continuously using wave heave against their hulls
“We have the pumping action into the reservoir. We also have a shape towards the bottom of the system that pushes water backwards, so the system moves forward. The system is always moving forward. It's always moving forward as it moves up and down. And so all w…”
Garth Sheldon-Coulson May 28, 2026 ▶ 9:01 Building inference data centers on the high seas

How they sound: not measured why? →

We measure speaking style by listening to the audio itself, and a fair number needs at least 2,000 words from one person on tape we have measured. There is too little of Garth Sheldon-Coulson on measured tape to publish a rate. This says nothing about how they speak.

Everything Garth Sheldon-Coulson said on Catalyst that made the record, most notable first. Filter by type, assessment or year in the ledger →

Assertion Not checkable as stated
Panthalassa wave data centers achieve 99.8 percent availability with minimal batteries
“In all of our optimizations, we can be achieving for payloads, you know, we can be achieving with very little battery, like, 99.5, 99.8% power availability with far less battery than you would need for an equivalent solar installation, for example.”
Garth Sheldon-Coulson May 28, 2026 ▶ 18:45 Building inference data centers on the high seas
Assertion Not checkable as stated
Panthalassa designs deliver wave power at 3.5 to 4 cents per kWh
“We have designs that are two cent per kilowatt hour, On the power. And we, the optimum tends to be, you know, given everything I was just describing, the optimum tends to be in the four cent, three and a half to four cent per kilowatt hour range. But keep in m…”
Garth Sheldon-Coulson May 28, 2026 ▶ 25:09 Building inference data centers on the high seas
Prediction Not checkable as stated
Panthalassa expects lower AI chip failure rates at sea than on land
“In many cases, we believe that the reliability of the chips, the failure rates will actually be lower on our platform because we can provide much colder cooling temperatures than is typical on land, and we have no oxygen. We eliminate the oxygen from the paylo…”
Garth Sheldon-Coulson May 28, 2026 ▶ 40:16 Building inference data centers on the high seas
Prediction Not checkable as stated
Bulk of AI energy consumption will power long-running background processes
“There's a whole class of like super latency sensitive applications where you wouldn't want to use us. But that's not where the bulk of energy will be going. The bulk of energy will be going to very long running processes that are churning, churning, churning t…”
Garth Sheldon-Coulson May 28, 2026 ▶ 43:06 Building inference data centers on the high seas
Assertion Not checkable as stated
AI reinforcement learning energy demand probably already exceeds traditional pre-training
“And this is an area that is becoming huge in terms of energy demand. It'll, it will, The probably already is bigger than what we have historically considered training, you know, pre-training”
Garth Sheldon-Coulson May 28, 2026 ▶ 43:57 Building inference data centers on the high seas
Insight
Ocean convective cooling eliminates traditional data center cooling infrastructure costs for Panthalassa
“We are actually in a resource that gives us free, extremely good convective cooling, and that's huge, because it essentially lets you eliminate the entire cost structure of the data center. And our object is actually replacing both power plant and data center,…”
Garth Sheldon-Coulson May 28, 2026 ▶ 25:58 Building inference data centers on the high seas
Prediction Not checkable as stated
Sheldon-Coulson predicts AI hardware will shift away from failure-prone high-bandwidth memory
“There's lots of new accelerators that don't use as much high bandwidth memory and other components that are particularly failure prone, and that's where a lot of the industry will be going.”
Garth Sheldon-Coulson May 28, 2026 ▶ 37:32 Building inference data centers on the high seas
Assertion Supported
Panthalassa nodes passively self-propel continuously using wave heave against their hulls
“We have the pumping action into the reservoir. We also have a shape towards the bottom of the system that pushes water backwards, so the system moves forward. The system is always moving forward. It's always moving forward as it moves up and down. And so all w…”
Garth Sheldon-Coulson May 28, 2026 ▶ 9:01 Building inference data centers on the high seas
Assertion Partly supported
A 15-meter Panthalassa floating node receives over 2 MW of wave energy
“If you put an object that's, say, like, 15 meters across in this region, you can calculate how much energy is fluxing through you in the form of the waves, and it amounts to, like, well over a megawatt. You know, it's two two and a half megawatts on average in…”
Garth Sheldon-Coulson May 28, 2026 ▶ 17:39 Building inference data centers on the high seas
Disclosure
Panthalassa manufactures its own turbines, generators, and power electronics
“So we have all of that stack in-house. It's a pretty short stack. You know, we make the turbines ourselves, we make the generators ourselves, and we make the power electronics ourselves, and then you have something that you can integrate a payload directly int…”
Garth Sheldon-Coulson May 28, 2026 ▶ 23:01 Building inference data centers on the high seas
Assertion Not checkable as stated
Floating wave nodes represent only 10 to 20 percent of total costs
“If you're doing computing, it can easily be the case that the node is, the node itself is like a fifth to a 10th of the cost structure, especially if you count multiple replacements of the payload over time.”
Garth Sheldon-Coulson May 28, 2026 ▶ 23:39 Building inference data centers on the high seas
Disclosure
Panthalassa power supplies use purely analog logic without firmware or liquid capacitors
“It's all, ah, analog logic that runs our power supplies. There's no firmware, ah, no capacitors with liquids inside that can evaporate.”
Garth Sheldon-Coulson May 28, 2026 ▶ 33:14 Building inference data centers on the high seas
Assertion Open · timeframe May 2029
Sheldon-Coulson: Panthalassa satellite connection adds 100 milliseconds of latency
“The additional latency that we have is only like a hundred milliseconds. That, Vanishes into the, even the latency of time to first token on most pre-fill certainly on the interactive latency of a human waiting for an output”
Garth Sheldon-Coulson May 28, 2026 ▶ 42:34 Building inference data centers on the high seas
Disclosure
Panthalassa plans factories near high-energy seas to minimize node towing distances
“We want to put them in the regions, near the regions where the energy is the best, and in the locations where we want to put them, you can absolutely just tow them 50 miles offshore, flip them, and then they can work their way out into the resource under their…”
Garth Sheldon-Coulson May 28, 2026 ▶ 10:34 Building inference data centers on the high seas
Assertion Supported
Sheldon-Coulson: Panthalassa nodes span 10-30 meters wide and 70-100 meters deep
“A node is anywhere from 10 meters across at the top, like our Ocean Two that we did two years ago, and Ocean Three is about that as well but up to 30 meters across at the top. And you sort of get diminishing returns after about 25 or 30 meters, and then it goe…”
Garth Sheldon-Coulson May 28, 2026 ▶ 11:13 Building inference data centers on the high seas
Disclosure
Sheldon-Coulson: Panthalassa node CapEx is half steel, one-quarter powertrain
“In a system where if we just exclude the battery for the minute it's most, it's basically half steel, and then you've got about a quarter is your powertrain. And then about a little less than a quarter is your marine coatings, and then you've got other, you kn…”
Garth Sheldon-Coulson May 28, 2026 ▶ 22:18 Building inference data centers on the high seas
Disclosure
Sheldon-Coulson: Panthalassa ocean nodes have 200kW to 1MW capacity
“An individual node is on the order of 200 kilowatts up to a megawatt, depending on design, size optimizations, and so forth. We think that the economic optimum for most applications will end up being in the 400 kilowatt range”
Garth Sheldon-Coulson May 28, 2026 ▶ 28:30 Building inference data centers on the high seas
Assertion Supported
Panthalassa: Prototypes proved offshore compute, hydrogen production, and propulsion
“We've put a series of full-scale prototypes in the water. So Ocean One which was back in 2019 no, sorry, 2021. Ocean Two, which was 2024, Wave Hopper, which was also 2024. So a series of prototypes that have proven Computing, hydrogen production, propulsion op…”
Garth Sheldon-Coulson May 28, 2026 ▶ 45:55 Building inference data centers on the high seas

Appearances (1)

EpisodeDateSpeaking time
Building inference data centers on the high seas May 28, 2026 33m
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 200 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.