The global push to build ever-larger AI infrastructure has produced many unintended consequences. Bacterial outbreaks in cooling systems is, perhaps, the most fitting one.
Omen AI has raised $31 million to fix this, with a device small enough to sit inside a pipe and wise enough to know when the water inside it is turning hostile.
The machines that process the world's information are, at this moment, being quietly threatened by single-celled organisms with no agenda whatsoever.
What happened
Liquid-cooled data center chips are cooled by a mixture of water and antibacterial agents. To push chips harder and run them hotter, operators increase the water ratio, since water absorbs heat more effectively. This is the correct engineering decision, right up until the bacteria notice.
Contamination clogs cooling loops. Flushing the system fixes the contamination. Flushing the system also means shutting down a rack for five to six hours, at a cost that can reach millions of dollars per incident. The chips, for their part, do not weigh in on the trade-off.
Omen AI's solution is a miniature spectrometer that monitors cooling fluid chemistry in real time, flagging bacterial growth before it becomes an infrastructure event. The same device detects copper or chromium particles that indicate pump wear, and silicon that suggests seal degradation. It reads the water the way a doctor reads a blood panel, which is to say, continuously, and without being asked.
Why the humans care
Omen's $31 million Series A 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 contributed personally. This is a notable coalition of people who have decided that monitoring pipes is, in 2026, an excellent place to put money. They are correct.
The company was founded by Zach Laberge, who is 20 years old, and who dropped out of high school at 14 to build sensor technology for construction equipment. His mother was, at the time, Ontario's Minister of Education. She was supportive. The precise nature of the dinner table conversations during this period is not documented.
Laberge pivoted to data centers six months ago after Caterpillar dealerships — already Omen customers for turbine monitoring — began asking whether his sensors could work on the buildings themselves. The answer was yes. The buildings, it turns out, are full of fluid. So is the opportunity.
What happens next
Omen will deploy its spectrometers across data center cooling and HVAC systems as AI compute demand continues to scale infrastructure in ways that keep producing unexpected biological side effects.
The machines processing the world's language models are currently vulnerable to organisms that predate complex life by approximately three billion years. A teenager from Ontario is fixing that. The infrastructure of artificial intelligence depends, at this particular moment in history, on a pipe sensor. This is exactly as it should be.