Omen AI raises $31M to monitor data center cooling fluid

A sensor company for data center plumbing is not the headline AI infrastructure usually gets, but Omen AI just closed a $31 million Series A on exactly that pitch, and the cap table is worth a second look. TechCrunch reports the round was led by Nava Ventures with CRV, Vanderbilt University, Mann+Hummel, Starhill Holdings, and Hard Launch Capital joining, plus executives from Bridgestone, GM, Johnson Controls, and TensorWave writing checks on the side. Total raised since the 2024 founding is $40 million.

What Omen actually sells is a spectrometer that sits in the coolant loop of a liquid-cooled chip system and reports what is happening chemically. As operators push toward more water-rich coolant mixes for better heat absorption, bacteria, metal wear and seal degradation that used to be invisible become real risks. The device looks for bacterial growth, traces of copper and chromium that indicate component wear, and silicon that suggests seals are going. Per the same reporting, the alternative today is a system flush of five to six hours at potential costs of millions in downtime. Founder Zach Laberge put the pitch plainly: "You're not risking huge amounts of downtime because you have no insight into what's going on chemically."

Why this matters past the funding press release is that liquid cooling is no longer a research project, it is what the next wave of AI compute is being built around, and the companies that own the picks and shovels around it tend to age well. Roughly a dozen data center customers are using Omen today, including TensorWave, the AMD-based AI compute cloud. That is a small but pointed customer base, because TensorWave is exactly the kind of independent operator that cannot field a Microsoft-scale facilities team and needs uptime monitoring as a bought product, not a built-in capability.

The honest caveats are the early-stage ones. A dozen customers is a thin base, and the reporting does not break out whether those are paid production deployments or pilots, what a spectrometer costs, or how it performs across different coolant chemistries. Laberge's previous startup, a construction equipment sensor company he started at 14, raised $3 million and shut down, which is a useful reminder that hardware in the field is harder than hardware in a deck. Take the specifics as reported, not settled.

The forward read is that some of the most interesting infrastructure bets right now are not the trillion-parameter model races, they are the quiet ones answering what is actually going to break first inside the buildings that run them. Coolant chemistry is a good candidate for that list, and if Mann+Hummel and the Johnson Controls executives on the cap table are signal, the existing filtration and HVAC supply chains read it the same way.