TL;DR: They're developing fuel cell systems (SOFC and r-SOC), aiming to deliver cleaner power on site for data centers, manufacturing campuses and utilities. They're taking on the usual limits in cost, power density and responsiveness.
Diving into the technical background — the best part- natural gas or hydrogen feeds the fuel cell, then that cell uses high temp + electrochemical reactions to convert the fuel to electricity, with no combustion needed (huge win). The idea is to make that electricity directly where it's needed, like at the data center, reducing dependence on grid expansion. Teragen also says the cells can work in reverse: using extra electricity to produce hydrogen, which can be stored and used as fuel later! (Small note: using natural gas still produces CO2, even without combustion.)
The founders are a big part of why this company stands out to me. CEO Ruofan Wang co-invented the core technology as a postdoctoral researcher at Lawrence Berkeley National Laboratory, and Teragen licenses it from the lab. He has a PhD in materials science and electrochemistry and has spent his career scaling energy technology. CTO Joshua Persky is a ceramic scientist with over 20 years of experience building fuel cell products and the teams manufacturing them. CCO Emerson Reiter started his career at PG&E on wholesale power markets and worked as a technical project lead at the National Renewable Energy Laboratory. All of this = very strong founder market fit.
The company spun out of Berkeley Lab and the Department of Energy's ARPA-E program, and recently closed a $6M pre-seed round co-led by BEVC and Energy Capital Ventures.
So why do I think they could win? Demand for electricity from data centers is growing fast. Making power on-site could reduce a facility's dependence on slow grid upgrades, though it still needs fuel and the approvals any on-site generation requires. Teragen is aiming to make its systems cheap enough to compete with gas turbines and quick to adjust their output, which matters because AI workloads fluctuate.
The tech is still in validation, moving from lab scale toward a commercial demonstration. Fuel cells for data centers also already exist. Bloom Energy, a public company, is deploying solid oxide fuel cells for customers like Oracle. Teragen isn't inventing the idea, so they have to prove their design is meaningfully cheaper and better than equipment customers can already buy.