A college dropout who spent his teenage years fixing bulldozer engines just raised $31 million to protect $200 billion worth of AI infrastructure.
He didn't build an AI company. He built a sensor for construction equipment. His customers turned it into an AI company, one odd question at a time.
What happened
Omen AI was founded in 2024 by Zach Laberge, who dropped out of high school at 14 to run his first startup — sensors for construction machinery, funded by $3 million that eventually ran out. Omen started with the same basic idea, aimed at industrial fleets: sensors that read the fluid running through a machine's system and flag wear before it causes a failure.
About six months ago, some of those industrial customers started asking a different question: could the same sensor work on the fluid systems inside data center buildings — the cooling loops keeping AI chips from overheating?
It turned out the underlying physics were nearly identical. On June 30, Omen closed a $31 million Series A led by Nava Ventures, taking its total funding to $41.5 million. It now monitors coolant health for data center operators representing $200 billion in assets and 10 to 14 gigawatts of capacity, alongside its original industrial customers.
Why it actually happened
Here's the part worth sitting with.
Omen didn't pivot because a board deck said "data centers are hot right now." It pivoted because customers who already trusted the product for one job started asking, unprompted, if it could do a second job.
That only worked because of something founders rarely stop to separate: the difference between your category and your capability.
Omen's category, on paper, was "construction equipment sensors." That's the label investors saw, that's the line on the business card, that's the market size someone put in a pitch deck.
But the actual capability underneath that label was narrower and far more valuable: reading a fluid's chemistry in real time to predict when a mechanical system is about to fail. That capability doesn't care whether the fluid is running through a bulldozer's hydraulics or a GPU rack's cooling loop. Metal wear looks like metal wear. Bacterial growth looks like bacterial growth. The physics of failure stayed the same — only the machine around it changed.
Most founders never make this distinction. They think their company is the market they picked on day one. So when that market gets crowded, or slow, or boring, they either grind through it out of loyalty to the original plan, or they panic-pivot into whatever's hot, dragging a capability that has nothing to do with the new category along with them.
Omen did neither. It kept the capability constant and let the label follow the customer.
The founder principle
Your company is not the market you picked. It's the capability you built to solve one specific version of a problem.
The market is a bet you made on day one, usually with less information than you have today. The capability — the thing you actually got good at — is the real asset. Markets change constantly. Capabilities transfer, if you notice they're transferable before you rename your whole company around whatever's currently declining.
This is also why some of the best pivots don't look like strategy at all. They look like a founder finally taking a strange customer question seriously, instead of filing it under "not our market."
The Strip-the-Label Test
A way to check whether you're sitting on a transferable capability without realizing it. Three questions:
Strip the label. Describe what you actually do without naming an industry. Not "we make construction sensors" — "we read fluid chemistry in real time to predict mechanical failure." If the second version is hard to write, you may not know what you're actually selling yet.
Who's pulling, not pushing? Has anyone outside your stated market ever asked, unprompted, whether your product could do something else? That's not a distraction from your roadmap. That's market research nobody billed you for.
Same physics, or same buzzword? Before chasing a new category, check whether it shares your original failure mode — or just shares a hot word. Omen's new market had the same physics: fluid degradation causing mechanical failure. A construction-sensor company chasing "AI" by bolting a chatbot onto its dashboard would just be chasing the buzzword.
Do this today
Write your category in one sentence. Then rewrite it with no industry name in it — just the underlying capability. If the second sentence is harder to write than the first, that's the gap to close before your next planning cycle.
Pull the last 90 days of inbound messages, support tickets, or odd sales questions. Look for anyone who asked if your product could do something outside its stated lane. Read it as free market research, not noise.
Before chasing whatever category is hot right now, check if its core failure mode actually matches something you already solve — not just whether it sounds adjacent.
Ask your best customers what specific capability they'd miss if you disappeared tomorrow — not what product category they'd say you're in. That answer is closer to your real moat than your own pitch deck.
If you're already mid-pivot, resist renaming the whole company around the new market on day one. Let the capability prove itself in the new context first. A young company can survive being known as "the fluid company" a while longer before it has to become "the data center company."
Closing thought
The best pivots rarely start in a board meeting. They start with a customer asking a question that doesn't fit the plan — and a founder who decides not to dismiss it.