The demand for compute power from AI is forcing data centers to squeeze more power out of every rack of GPUs. Is it one result? Bacterial outbreak.
The liquid in a water-cooled chip is a mixture of water and a substance that inhibits bacterial growth. To make chips run at higher temperatures, data center managers can change the formulation to contain more water. This allows for better heat absorption, but creates nasty contamination that can clog the flow. To resolve this, flash your system. This means shutting down the rack for 5-6 hours and can cost millions of dollars.
Omen AI has a solution. It’s a small spectrometer that can monitor the status of body fluids in real time and spot bacterial growth before it becomes a big problem. “We don’t have any insight into what’s going on chemically, so we’re not at risk for huge amounts of downtime,” CEO and founder Zach Laberge explains.
Today, Omen AI announced that it has raised $31 million in a Series A round led by Nava Ventures with participation from CRV, Vanderbilt University, Mann + Hummel, Starhill Holdings, Hard Launch Capital, and private investments from executives from Bridgestone, GM, Johnson Controls, and TensorWave.
Laberge founded his first company in 2020 at the age of 14, raising $3 million to install sensors on construction equipment, and ultimately dropping out of high school. (His father and mother, a former Ontario education minister, were supportive of his plan to forge his own path.)
After that startup shut down, Laberge launched Omen in 2024 with the idea of focusing on the key fluid systems that make construction equipment smart enough to know when it needs repair. The idea was to replace the time-consuming process of extracting samples and sending them to the lab with real-time recognition. In addition to bacterial growth, the device can detect pump or pump wear if copper or chromium is detected, or seals if silicone is detected.
Caterpillar dealerships were Omen’s major early customers for its heavy-duty vehicle business, but Cat is also a major supplier of gas-powered turbines and generators that provide on-premises power to data centers. It didn’t take long for Omen to figure out where the wind was blowing.
“It was kind of a transition,” Laberge told TechCrunch. About six months ago, “a lot of dealers were saying, ‘We’re starting to put sensors on our turbines, but is there anything we can do on the building side?'”
Omen discovered that these buildings were filled with liquids, from the HVAC systems to the cooling of the chips. Omen discovered a new and rapidly growing group of potential customers and began focusing on data centers.
“It’s rare to see such a young founder respected by established, large companies in a sector that’s a little more slow-moving,” said Corey Leras, a partner at Nava Ventures who is a member of Omen’s board of directors. “Particularly in the case of Omen, much of our enthusiasm came through introductions with large customers that quickly validated their approach.”
Omen, which has raised $40 million since its founding in 2024, is building its services in collaboration with more than a dozen data center customers, including TensorWave, a company building an AI computing cloud on AMD chips.
“The fluids that flow through these large systems are critical variables that most of the industry is blind to,” TensorWave president Piotr Tomasik said in a statement. “We see the future of infrastructure coming, just the way we do: improving monitoring to best support our compute customers.”
While many organizations rely on mailing liquid samples to laboratories to gain insights, Omen is not alone in developing on-premise analysis. Pyxis, an established water monitoring company, introduced a data center coolant monitoring product earlier this month.
The major technological advances that have made this approach possible are recent improvements in both optical technology and signal processing software. “The hardware is cheap enough that it makes sense to play at scale, and then signal processing allows us to make more sense out of the noise,” Laberge says.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
