Data centers run hot, and cooling them is turning into one of the strangest gold rushes of the AI boom. Omen AI, a startup building real-time sensors for the fluids that keep artificial intelligence chill, has acquired competitor Aquatic Labs in a $16 million cash-and-stock deal that says less about rivalry and more about how fast this market is moving.
The merger brings together two complementary ways of reading water chemistry, one born in heavy machinery and one born in ocean science. Here is what the deal tells us about the business of keeping AI from overheating, and why the people building it are not remotely worried about the cracks forming in the AI spending boom.
The Omen AI Deal at a Glance
Omen AI was founded in 2024 with a narrow, unglamorous mission: put a real-time sensor inside the cooling loops of data centers so operators can see, moment by moment, whether the fluid circulating over their chips is still healthy. On November 2, 2026, the company announced it is buying Aquatic Labs, a rival sensor startup with a different but compatible instrument, in a deal worth $16 million in cash and stock.
A merger introduced by an investor
The deal did not grow out of a bidding war. Omen AI founder Zach Laberge was introduced to Aquatic Labs chief executive Allan Adams by an investor in the filtration space who saw two small companies building adjacent halves of the same product. Rather than spend years courting the same data center customers from opposite ends of the chemistry problem, they agreed to build one instrument together.
What each side brings
Omen AI’s sensors look for elements in the water: traces that betray failing valves or bacterial growth threatening to choke off flow. Aquatic Labs’ instrument tests alkalinity, the acidity balance that decides whether a cooling loop is quietly corroding from the inside. Combined, Adams told TechCrunch, the two technologies create a “much richer picture of the water and the fluid” than either could produce alone.
Deal snapshot
| Item | Detail |
|---|---|
| Acquirer | Omen AI, founded 2024, raised a $31 million Series A in June 2026 |
| Target | Aquatic Labs, real-time water alkalinity sensing |
| Price | $16 million, cash and stock |
| Key role | Aquatic Labs CEO Allan Adams joins Omen AI as chief science officer |
| Combined product | One integrated sensor expected on the market by early 2027 |
Why $16 million is the telling number
A $16 million price for a sensor startup with data center customers reads as modest next to the nine-figure rounds the AI infrastructure market routinely prints. That gap is the opportunity in miniature: the computing layer of this market attracts the spectacle capital, while the physical layer trades at industrial multiples. Omen AI is buying measurement capability cheaply because, until recently, almost nobody priced it as strategic.
Expect the comparison to collapse within a cycle. Water is now a board-level constraint for hyperscalers, and instruments that govern water use will be priced accordingly. The Omen AI deal may look, in hindsight, like the moment the fluid layer got its first real valuation.
Why Water Is the Real AI Constraint
The acquisition lands against a backdrop that would have sounded absurd five years ago: water is becoming a hard constraint on artificial intelligence. United States data centers use somewhere between 17 and 19 billion gallons of water a year to cool their hot-running chips, and The Water Foundation expects that number to triple within the next five years.
Billions of dollars, disappearing gallons
For investors pouring billions into the challenge of scaling AI, that trajectory is expensive and politically tricky. Water rights are contested territory in much of the country, and our report on how AI has come for Arizona’s water shows what happens when hyperscalers and farmers reach for the same aquifer. Every gallon a cooling loop keeps out of the drain is a small political victory as much as a financial one.
The slow ritual of sampling
One way to use less water is to monitor the cooling fluid better, a task that today reads like laboratory ritual. Operators send samples away to third-party labs, wait days for results, and act on chemistry that has already changed. During that waiting window, valves corrode, bacteria colonize, and operators fly blind. Omen AI was built to collapse that multi-day process into a live signal.
Efficiency metrics are not enough
The industry has tried to manage thirst through aggregate metrics; as we covered, water usage effectiveness has improved at AI data centers even as total consumption climbs. But averages hide the failure modes. A loop can post a perfect efficiency score while slowly poisoning itself with alkaline drift. Point-level chemistry, measured continuously, is the layer beneath every efficiency number, and almost nobody instrumented it. That gap is the market Omen AI and Aquatic Labs created.
A pattern seen across critical infrastructure
Every mature infrastructure sector eventually grows a nervous system. Power grids installed phasor measurement units, aircraft fleets grew telemetry, and shipping containerized its cargo data. The data center industry is repeating the sequence at compression speed, and cooling fluid is the last unmeasured major variable. Omen AI is positioning for the moment operators decide that flying blind on chemistry is simply bad management.
Two Sensors, One Picture: How the Technologies Combine
The two startups were, in Laberge’s telling, classic competitors: same customers, same problem, different answers. Put side by side, their instruments read complementary halves of the water’s story.
| Dimension | Omen AI sensor | Aquatic Labs sensor |
|---|---|---|
| What it reads | Elements present in the water | Alkalinity of the water |
| What that reveals | Failing valves, bacterial growth choking flow | Corrosive chemistry drift inside the loop |
| Origin story | Hydraulic fluid analysis for construction machines | Ocean sensing from MIT and Woods Hole research |
| After the deal | Core of the merged platform | Integrated module; CEO becomes chief science officer |
From two boxes to one instrument
The engineering goal is embarrassingly simple to state: one sensor, both readings, no laboratory. The combined teams are working to fold both sensing technologies into a single housing, and they expect the merged instrument on the market by early next year. Until then, data centers that want the full picture buy two instruments and stitch the dashboards themselves.
Why alkalinity matters more than it sounds
Alkalinity is an unsexy word for an expensive problem. Cooling water that drifts too alkaline eats metal from the inside; too acidic and it deposits scale that insulates chips from the very cooling meant to save them. Neither failure announces itself. By the time symptoms appear, operators are replacing hardware or flushing loops, exactly the expensive, wasteful outcomes continuous measurement is meant to prevent.
Automating what humans eyeballed
Both founders framed the merger around a shared vision of automation: replacing a slow, manual, human-judgment process with an always-on measurement that triggers action before damage compounds. The chemistry was never mysterious; the missing piece was presence. Omen AI’s bet is that once operators see their loops in real time, they will never again accept a five-day-old lab report as the truth.
The quiet economics of calibration
One understated prize in deals like this is calibration data. Sensors earn trust by agreeing with laboratory results thousands of times, and the merged dataset will teach Omen AI how element signatures and alkalinity interact across hundreds of real cooling loops. That accumulated knowledge is the moat competitors cannot buy their way around, because it accrues only to whoever is actually installed.
The Founders Behind the Merger
Every acquisition is a story about two temperaments, and this one offers a textbook contrast. Laberge dropped out of high school to pursue his first company, which did fluid analysis for the large construction machines that depend on hydraulics. He learned cooling loops the way mechanics learn engines: by watching what happens when nobody maintains them.
The mechanic and the physicist
Adams is the mirror image: trained as a physicist, he spent much of his career developing advanced ocean sensing technology as an MIT professor and Woods Hole Oceanographic Institution researcher before founding Aquatic Labs. The sea, it turns out, is just a very large cooling loop with worse plumbing. Instrumentation that can read chemistry in the open ocean translates gracefully to instrument-grade readings inside a pipe.
Why this pairing suits the market
Data center buyers are famously skeptical, and the combined pedigree answers their two objections at once. Laberge brings scar tissue from industrial environments where a bad reading costs a machine; Adams brings scientific credibility where chemistry claims must survive peer review. Omen AI gets the rare ability to sell simultaneously to the maintenance trailer and the laboratory.
One team, one roadmap
Adams joins Omen AI as chief science officer, keeping the Aquatic Labs technology alive inside the merged roadmap rather than parking it in a graveyard of acquired startups. That is a small but instructive data point about how this corner of the AI infrastructure market is consolidating: not absorption and shutdown, but splicing, because the remaining problems are integration problems.
The Economics of Not Dumping the Water
The value Omen AI sells is not water savings, exactly. It is the right to keep running. When a fluid system gets out of whack, data center operators typically choose to dump the water rather than risk damage to expensive equipment, which means shutting operations down until the loop is restored.
Downtime is the real bill
Every paused rack is revenue evaporating at the precise point where chips are most valuable. Monitoring flips the calculation: operators can use the same fluid for longer, treat problems chemically instead of surgically, and minimize the costs of a clean-up. The sensor pays for itself the first time it talks an operator out of an emergency flush.
The scale Omen AI is watching
Omen AI’s initial sensors will monitor chips pulling in 250 to 300 megawatts of power by the end of the year, worth about $5 to $7 billion of compute, Laberge told TechCrunch. The company’s stated ambition is the equivalent of one gigawatt of chips under monitoring by the end of 2027.
Why operators buy after the fact
Sensor companies in critical infrastructure face a brutal sales truth: buyers move fast after failures and slow before them. Omen AI’s pitch is the rare one where the purchase price is trivial against the asset being protected. A few thousand dollars of instrumentation guarding tens of millions of dollars of silicon is not an innovation budget decision; it is housekeeping.
The arithmetic improves with density. A single hyperscale campus may run hundreds of parallel cooling loops, each a candidate measurement point, which means the per-site revenue ceiling rises automatically as new halls come online. Omen AI’s one-gigawatt goal is best read as an announcement that the company has seen the pipeline of loops waiting to be instrumented, not a wish cast into the market.
The insurance framing
Ask any insurer about critical equipment and you hear the same lesson restated: prevention technology sells slowly until the first catastrophic claim, then sells instantly. Omen AI is positioning inside that dynamic. The industry’s growing library of cooling failure stories, most of which never leave the affected campus, is a marketing budget the company does not have to spend.
The broader economic point stands with or without a bubble: at any plausible level of AI investment, cooling fluid management remains a rounding-error cost with failure costs in the millions. Sensors priced accordingly will find buyers in expansion or contraction alike.
Bubble Worries and a Defiant Answer
Ask a hard infrastructure question long enough and the AI bubble conversation arrives whether you invited it or not. AI companies once estimated that as much as $4 trillion would be spent on AI infrastructure by 2030, and some investors have grown openly skeptical of the buildout, with hyperscaler stocks plummeting in recent days.
The view from the cooling loop
Omen AI’s team is not worried about shifting sentiment, and their reasoning is structural: their work happens after a data center has been built. The sensors attach to assets that already exist, financed by commitments already signed. A cooling sensor company is downstream of the buildout’s sunk costs, not upstream of its promises.
“Still the biggest market in human history”
“We’re working customers on stuff they’re building today or have already built, and so it doesn’t really affect us because all of those projects were financed 12 months ago,” Laberge said of the spending doubts. His larger claim is a thesis in miniature: even if you sliced the AI market in half, he argues, “this is still going to be the biggest market in human history.”
Why the skeptics and the builders can both be right
The tension between the spending skeptics and the Omen AI camp may be less a disagreement about AI than about time horizons. If the buildout slows, the installed base still needs cooling, and cooling gets more scrutiny, not less, as capital discipline tightens; our piece on using AI to mitigate data centers’ environmental threat makes the same case from the energy side. Efficiency plays win in boom and bust alike, which is precisely why the giants pay attention to startups in this niche.
A market with policy written around it
Measurement companies enjoy a peculiar advantage in regulated markets: rules that demand reporting create customers by decree. Water discharge permits, efficiency disclosures and community impact reviews all generate demand for exactly the data continuous sensors produce. Omen AI sells into that demand without waiting for it to become economic, which gives the category a floor most AI infrastructure plays never had.
What the Omen AI Acquisition Means for Thirsty AI
The Omen AI acquisition of Aquatic Labs is small in dollar terms and large in signal. It says the water-sensing market has matured past the point where one startup per continent survives on vibes, and into an era where integration, breadth of measurement, and installed footprint decide the winner.
Consolidation is the maturity tell
When competitors start merging instead of competing for the same pilot programs, the category has crossed a threshold. Regulators are pushing the same direction from the other side: California has tightened its rules on data center energy and water use, and the pattern of measurement-first compliance favors companies that already produce continuous, auditable fluid data.
Latecomers should take notes from this playbook. Building a sensor is a physics problem, and physics gets commoditized; owning the installed base of calibration history is an operational problem, and operations compound. That is why Omen AI paid cash and stock for reach rather than waiting to earn it organically, and why the next deal in this niche will probably be larger than this one.
Outside the loop, the same physics
Adams tells TechCrunch he hopes the technology reaches well beyond data centers, with a reminder that cooling water does not stay inside the building: “water goes into data centers, and then it comes out and gets dumped into municipal water, and you want that whole process to happen in a way that’s as safe and responsible as possible — and that means you got to measure stuff.” Municipal discharge, industrial reuse, district heating: every one is a measurement market wearing different clothes.
The quiet layer of the AI stack
The AI infrastructure story is usually told in chips, gigawatts and billion-dollar deals. The Omen AI acquisition tells a quieter version: the future of artificial intelligence may hinge on boring instruments that read pH, alkalinity and trace metals in pipes. If the trillion-dollar buildout behaves like every industrial era before it, the durable fortunes will belong to whoever understands the fluids.
What to watch next
Three signals will tell the story of this category over the next year. First, whether Omen AI ships the integrated sensor on schedule, because roadmap credibility decides enterprise adoption faster than any feature. Second, whether a hyperscaler installs the technology across its fleet, turning a supplier into a standard. Third, whether municipal water authorities begin demanding the same measurement data Omen AI produces, which would convert a private efficiency tool into licensed public infrastructure.
References
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
