A bacterial outbreak quietly brewing inside a liquid-cooled server rack might sound like a niche engineering headache — but for data center operators running AI workloads around the clock, it translates directly into millions of dollars in downtime. That is exactly the problem Omen AI is tackling with a new approach to data center fluid monitoring, and the startup just secured $31 million in Series A funding to scale it fast.
Summary
Key takeaways
- Omen AI built a miniature spectrometer that monitors liquid cooling fluid health in real time, detecting bacterial contamination before it forces a rack shutdown.
- Bacterial contamination in coolant systems can force a data center rack offline for five or six hours, at a potential cost of millions of dollars per incident.
- The $31 million Series A was led by Nava Ventures, with participation from CRV, Vanderbilt University, Mann+Hummel, Starhill Holdings, Hard Launch Capital, and executives from Bridgestone, GM, Johnson Controls, and TensorWave.
- Omen has raised $40 million total since its 2024 founding and currently serves about a dozen data center clients, including TensorWave.
- Competitor Pyxis rolled out a comparable coolant monitoring product earlier this month, signaling growing industry attention to the problem.
The Hidden Chemistry Problem Inside Liquid-Cooled Data Centers
Liquid cooling is no longer optional for high-density AI infrastructure — it is becoming the baseline. But the fluid that runs through these systems is more chemically delicate than most operators realize.
The coolant is typically a mixture of water and a bacteria-inhibiting substance. The trade-off is straightforward: increase the water content and you improve heat absorption, which lets chips run hotter and harder. But more water also creates a more hospitable environment for bacterial growth. Left undetected, that contamination clogs the coolant flow, and the only fix is a full system flush.
That flush is expensive. Shutting down a single rack to clear a contaminated cooling loop can take five or six hours — time during which compute capacity is simply gone. For operators running AI inference or training workloads, that is not an inconvenience. It is a serious financial hit.
Flying Blind on Fluid Chemistry
Until now, most data centers have handled coolant health the same way for decades: extract a fluid sample, mail it to a laboratory, and wait for results. By the time the lab report comes back, a contamination problem may already be well advanced.
“The fluid running through these massive systems is a critical variable that most of the industry is flying blind on,” said Piotr Tomasik, president of TensorWave, one of Omen’s current customers. TensorWave builds AI compute clouds on AMD chips and has become a reference customer for Omen’s approach.
The gap between when a problem starts and when a lab result arrives is precisely where Omen AI is planting its flag.
Omen AI’s Spectrometer and What Makes It Different
The core of Omen’s product is a compact spectrometer installed directly within the fluid system — no sample extraction, no shipping delay, no waiting. It continuously reads the chemical composition of the coolant in real time, flagging bacterial growth early enough for operators to act before a shutdown becomes inevitable.
Beyond bacterial contamination, the device can also detect wear in pumps by spotting traces of copper or chromium in the fluid, and identify seal degradation through silicon particles. That gives data center operators a much broader window into the health of their cooling infrastructure than any periodic lab test could provide.
As CEO and founder Zach Laberge put it: “You’re not risking huge amounts of downtime because you have no insight into what’s going on chemically.”
What Made the Technology Viable Now
The timing of Omen’s approach is not accidental. Two converging developments made a miniaturized, affordable, on-premises spectrometer feasible: recent advances in optical technologies and improvements in signal processing software.
“Hardware is just cheap enough that it makes sense to play at scale, and then signal processing lets us make more sense out of the noise,” Laberge explained. Without both of those ingredients, building a device small and inexpensive enough to deploy across dozens of racks would not have been economically realistic.
That technological unlock matters beyond Omen’s own story. It helps explain why the real-time coolant analytics space is heating up more broadly — Pyxis, an established water-monitoring company, launched its own data center coolant monitoring product just this month. The convergence of optical hardware costs and software capability appears to be opening the market to multiple entrants simultaneously.
Company Growth, Funding, and Strategic Partnerships
Omen AI’s path to data centers was not a straight line. Zach Laberge founded his first company in 2020 at age 14, raising $3 million to install sensors on construction equipment — and famously dropping out of high school to do it, with his parents’ support. After that startup wound down, he launched Omen in 2024 with a broader vision focused on fluid systems as a diagnostic layer for industrial machinery.
From Caterpillar Dealerships to Data Centers
The pivot to data centers was driven by the company’s existing customers. Caterpillar dealerships were among Omen’s earliest clients in the heavy vehicles segment. Caterpillar is also a major supplier of gas-powered turbines and generators used to power data centers on-premises, which put Omen in direct contact with operators managing large-scale building infrastructure.
About six months ago, dealerships began asking whether Omen’s sensors could be applied to the building side — the turbines, HVAC systems, and chip cooling loops running inside data center facilities. Omen quickly recognized that those buildings were full of fluid systems that needed exactly the kind of monitoring it had already built for construction equipment.
The $31 Million Round and Who Backed It
The Series A round of $31 million was led by Nava Ventures, with participation from CRV, Vanderbilt University, Mann+Hummel, Starhill Holdings, and Hard Launch Capital. Executives from Bridgestone, GM, Johnson Controls, and TensorWave also invested personally. Combined with earlier capital, Omen has raised $40 million since its founding.
Cory Rellas, a partner at Nava Ventures who now sits on Omen’s board, offered a notable signal about how the round came together: “For Omen in particular, much of our diligence came through our introductions with large customers which quickly validated their approach.” That investor-customer validation loop — where enterprise buyers effectively de-risk the investment thesis — is a strong signal of real commercial traction, not just technical promise.
Omen currently works with approximately a dozen data center customers as it builds out its product offering. TensorWave is the most prominent named client, but the breadth of investor participation — spanning automotive, industrial, and compute infrastructure executives — suggests the company is already bridging multiple sectors.
Why This Moment Matters for AI Infrastructure
The pressure to run AI chips hotter and more efficiently is not easing. As demand for compute scales, data center operators will keep pushing cooling systems closer to their chemical and mechanical limits. That pressure makes the trade-off between heat absorption and contamination risk a permanent engineering constraint, not a temporary one.
Real-time fluid monitoring addresses a gap that has largely been invisible in the infrastructure conversation. Compute hardware, power delivery, and network bandwidth get enormous attention. The chemistry of the coolant running through the system gets almost none — even though a single contamination event can idle a rack for the better part of a working day. Omen’s bet is that as liquid cooling becomes ubiquitous in AI data centers, chemical visibility will become as standard a requirement as uptime monitoring or power management.
With Pyxis entering the same space at roughly the same time, the competitive dynamics will likely accelerate both product development and customer awareness. For data center operators still relying on periodic lab tests, the question is no longer whether real-time coolant analytics will become standard practice — but which vendor gets embedded in their infrastructure first.
FAQ
What problem does Omen AI’s spectrometer solve for data centers?
It monitors liquid cooling fluid health in real time using a miniature on-premises spectrometer, detecting bacterial contamination early enough to prevent costly rack shutdowns that can last five or six hours.
Why is contamination a risk in liquid cooling systems at data centers?
Increasing the water content in cooling fluid improves heat absorption and lets chips run hotter, but it also creates conditions favorable to bacterial growth. That contamination can clog coolant flow and force operators to flush and shut down entire racks.
Who are some of Omen AI’s key partners and customers?
Omen AI’s $31 million Series A round included investors such as Nava Ventures and CRV, along with personal investments from executives at Bridgestone, GM, Johnson Controls, and TensorWave. Its customer base includes Caterpillar dealerships and TensorWave, which builds AI compute clouds on AMD chips.
How does Omen AI’s technology improve over traditional fluid testing methods?
Traditional methods require operators to extract coolant samples and send them to an external laboratory, introducing significant delays. Omen’s spectrometer provides continuous, real-time chemical analysis directly within the fluid system, eliminating the lag between a problem developing and an operator being alerted.
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.
