Data center coolant monitoring startup Omen AI Inc. is trying to fix one of the most pressing yet little-known challenges in the artificial intelligence industry after raising $31 million in Series A funding today.

The round was led by Nava Ventures and saw participation from CRV, Vanderbilt University, Mann+Hummel, Starhill Holdings and Hard Launch Capital. Executives from Bridgestone Inc., TensorWave Inc. and General Motors Co. also invested in the round.

The problem Omen AI is trying to tackle involves bacteria, which surprisingly can cause significant problems for liquid-cooled AI chips. As data centers try to squeeze more and more graphics processing units into each rack to boost cluster sizes, those chips run increasingly hotter. But the fluid that’s used to prevent those chips from cooking becomes more welcoming to bacteria the hotter those chips become.

In an interview with TechCrunch, Omen AI founder and Chief Executive Zach Laberge explained that data centers use a liquid coolant that’s made up of water and an additive that’s meant to suppress the growth of bacteria. However, sometimes it’s necessary to add a bit more water to that mixture so that the chips can be pushed even harder. More water allows the liquid to absorb more heat, but the wetter the mix is, the more likely it is to become contaminated – not only with bacteria, but also with tiny, sometimes microscopic pieces of metal and other particles. The problem is that these can eventually prevent the liquid from flowing through the system as it should.

The standard fix is to occasionally flush the system, dump the coolant and replace it with a fresh mix. But doing this means taking an entire rack of servers offline for five or six hours at a time, which can cost data center operators millions of dollars. Omen AI offers an alternative with a system that can continuously monitor the health of the coolant and spot problems with it before a flush becomes necessary. “You’re not risking huge amounts of downtime because you have no insight into what’s going on chemically,” Laberge explained.

At just 21 years old, Laberge is an incredibly young age for an infrastructure company founder. But he doesn’t lack experience. He founded his first business in 2020 while still 14, raising $3 million to integrate sensors with construction machinery so those machines could be better maintained. Laberge soon dropped out of school to run that company, with the full backing of his parents.

He started Omen in 2024, and originally targeted the same industry with the idea of building a system that could monitor the fluid in their pneumatics, so that operators don’t have to constantly take samples and send them to a laboratory for analysis. One of Omen’s customers is the construction and mining equipment manufacturer Caterpillar Inc., which also supplies turbines and generators for on-site power generation at data centers. When the AI boom sparked a massive data center buildout across the world, some of Caterpillar’s dealers became heavily involved in those construction projects, and eventually asked Omen if it’s also possible for them to monitor the buildings, too.

That was when Laberge realized that data centers were so reliant on fluid, in everything from their HVAC systems to the chip cooling systems. “Taking a sample, shipping it to a lab, and waiting days for results is dangerously inadequate when you’re protecting billions in GPU infrastructure and operating industrial machines,” he said. “Omen AI was built to prevent catastrophic failure. We help data centers push their hardware to the absolute limit, unlocking compute performance operators didn’t know they had.”

The company now has about a dozen data center operators as customers on its books, including Tensorwave, the neocloud that focuses exclusively on Advanced Micro Devices Inc.’s GPUs. Those customers can choose between rigging up a permanent sensor array that connects directly to a server rack’s fluid system to continuously track metal content, bio contamination and wear patterns, or a portable diagnostic unit that can be brought to any machine for an immediate diagnosis. In both cases, they monitor coolant for more than 21 elemental signatures, replacing the old “sample-and-wait” model with continuous, real-time intelligence.

For Omen AI, this is a growing market. AI data center rack densities have increased well beyond what air-based cooling system can handle, which is why investors are increasingly looking to fund startups that can enhance the efficiency of liquid cooling systems. That’s why Iceotope was able to raise $26 million in Series B funding in May. Omen AI faces another, more direct competitor in Pyxis Lab Inc., which last month launched its own coolant chemistry monitoring system.

Cory Rellas of Nava Ventures said he’s betting on Omen AI because the cost of unplanned failures in data centers can be staggering, running into tens of millions of dollars. “Despite the high stakes, these systems are still monitored with lab tests that take days,” Rellas said. “Omen AI built the solution: continuous, real-time visibility into the health of the machines doing the world’s most critical work.”

Image: Omen AI