Omen AI raises $31M for real-time data center coolant monitoring

A 21-year-old founder pitching a spectrometer for cooling fluid isn't where I expected this week's AI infrastructure story to land, but TechCrunch reports that Omen AI just closed a $31 million Series A led by Nava Ventures, with CRV, Vanderbilt University, Mann+Hummel, Starhill Holdings, and Hard Launch Capital joining the round. Personal cheques came in from executives at Bridgestone, GM, Johnson Controls, and Tensorwave, which is itself an Omen customer building an AI compute cloud on AMD chips.

The premise is unglamorous in a useful way. Liquid-cooled chips run on a mix of water and an additive that suppresses bacteria. Operators who want more heat absorption dial the water up, which works until the mix gets wet enough to invite contamination that clogs the flow. The fix, according to the reporting, is to flush the system, which can mean taking a rack offline for five or six hours at a cost that runs into the millions. Omen's answer is a small spectrometer that reads fluid chemistry continuously and flags trouble early, picking up bacterial growth, copper or chromium that points to pump wear, and silicon that points to seal degradation.

Why this matters is the shape of the AI buildout right now. The marginal GPU rack is becoming the binding constraint on a lot of model work, and an unplanned six-hour outage isn't an ops nuisance anymore, it's missed training runs and broken SLAs. Founder and chief executive Zach Laberge frames the pitch as the difference between flying blind and, in his words, having "insight into what's going on chemically." With about a dozen data center customers signed including Tensorwave, that pitch seems to be landing.

The honest caveat is that this is one report on a young company. The customer count is small, total funding sits at roughly $40 million since the 2024 founding, and Pyxis, an established water-monitoring firm, has reportedly just rolled out a competing data center coolant product, so the moat is narrower than the headline implies. What the reporting doesn't give you is unit economics, false-positive rates on the spectrometer, or how Omen prices against the flush event it claims to prevent. The interesting thread to pull is whether chip vendor reference designs start specifying continuous fluid sensing as standard. If they do, picks-and-shovels companies like Omen are suddenly playing in a much bigger lane than a dozen pilot accounts.