TLDR:
Comfort Systems USA — a company that installs HVAC and wires up buildings — is up around 290% in a year, because a data centre is mostly a very expensive plumbing problem and someone has to build it.
The category (technical services) ticks every box serial acquirers love: ~$700bn in the US alone, fragmented, essential, recurring, and now riding the best capex tailwinds going.
The interesting question isn't whether it can be rolled up — private equity sorted that out years ago. It's who can do it without blowing themselves up, and whether there's a version of this for the rest of us. (Spoiler: there is, and it's in Europe.)
Why is an air-conditioning company up 290%?
Over the last twelve months, a company that installs air conditioning and wires up switchgear has gone up around 290% (in the year to 18 June 2026).
I'll admit my first reaction was that someone had mislabelled a chart. HVAC contractors are not supposed to trade like semiconductor stocks. To be clear, this isn't the market suddenly discovering that technical services can be rolled up — private equity has known that for years. What's new is narrower: an unglamorous MEP contractor (mechanical, electrical, plumbing) has become a public proxy for the AI build-out, and the market is now pricing it like one. Strip away the GPUs and a hyperscale data centre is, to a first approximation, an enormous and very fussy plumbing problem. Someone has to build that. Increasingly, it's Comfort Systems.
So this is a piece about technical services: the hard side of facilities. HVAC, electrical, plumbing, fire and life safety, lifts, and security systems — everything in a building that keeps it open, running and compliant.
It is not a glamorous market. It is, however, exactly the kind serial acquirers fall for, because it ticks every box at once:
Enormous, and nobody's in charge. Comfort puts US commercial, industrial and institutional mechanical and electrical contracting at roughly $700bn a year — and that's the US alone. The largest player in HVAC holds under 2% share. In other words, a vast market that is ripe for consolidation.
Not optional. Buildings can't open, operate or stay compliant without it. Demand is written into building codes, safety rules and insurance requirements — mandated, not discretionary.
The install seeds the annuity. Win the project, get close to the customer, and you're first in line for the years of service, repair, upgrade and replacement that follow. The installed base is the recurring revenue.
Hard, which is the moat. Labour is scarce and execution separates the good operators from the bad. Complexity keeps the market inefficient — and inefficient markets are where consolidators make money.
The macro is, for once, on side. Data centres, reshoring, electrification, power-load upgrades, efficiency retrofits. The tailwinds are about as strong as they get.
The interesting question, then, isn't whether technical services can be rolled up — that's settled. It's who can do it in the right part of the market without blowing themselves up. Comfort Systems is the cleanest case study for both halves of that sentence.
Is it real? How Comfort actually did it
There are two playbooks here, and the gap between them is the whole point.
The old one is the textbook roll-up. Formed in 1997 by bolting a group of mechanical contractors into a single public company, then buying complementary businesses with good crews, real technical know-how and strong local positions. Today that's 50 operating units across 190 locations and ~22,700 employees.
The new one is the bit worth paying attention to. Comfort stopped being just a buyer of local MEP shops and built genuine capability — design/build, prefabrication, and modular and off-site construction. The things that matter when a job is too technical to win on price. The point wasn't scale for its own sake; it was using scale to move into the most technical, highest-value work, where you compete on capability rather than price.
You can see the shift in where the revenue now comes from. Technology — overwhelmingly data-centre hyperscalers — is now 45% of revenue, and manufacturing another 22%. So two-thirds of the book sits in data centres and industrial build-outs: the densest, most technical, least-commoditised corner of the market. That re-mix, more than anything else, is what the share price has been reacting to. And this is booked work, not a hopeful pipeline — the backlog nearly doubled in a year, to ~$12bn.
How the money's made. Two things are true at once:
It's still a project business. New construction alone is 63% of FY25 revenue, with more on top from construction and renovation in existing buildings. Projects pay the bills today.
Service is the long game. Maintenance, monitoring and repair are only ~13% of revenue today — but that's the annuity: once a system's installed, you're first in line when it needs servicing, upgrading or replacing.
And running through both: the better businesses sell reliability, design judgment and speed rather than labour hours — capability that's worth far more in a data centre than a generic office fit-out.
The engine: multiple arbitrage. Then there's the bit that makes consolidation actually pay. Buy something small and cheap, fold it into something big and expensive, and you've created value before changing a single thing operationally:
The short version: a $1m-EBITDA contractor bought at 7x ($7m) is worth $12m inside a platform valued at 12x — $5m created on the spread alone, before you've touched the business — and the gap widens to ~$9m once you nudge margins from 15% to 20% (taking EBITDA to ~$1.3m).
It's never quite that clean in practice — the spread only turns into real money if you can actually integrate and run the thing. But that, in a sentence, is why anyone consolidates a fragmented market. Comfort ran exactly that play, and then made the decision that matters more: it pointed all that consolidation at the densest, most technical work — data centres, advanced manufacturing — where the job is too hard to win on price alone. That's why it now trades like an infrastructure compounder rather than a regional HVAC shop.
Could you actually do this?
Here's where it gets relevant for the rest of us — because Comfort Systems is now a ~$69bn company and not, you'll be amazed to hear, available to a first-time searcher. But the playbook is, and there's more than one way to run it.
There's no single template. The three best-known plays sit on a spectrum — from the hardest, highest-value work down to the lightest and most commoditised:
Approach | Example | The bet | The catch |
|---|---|---|---|
Own crews, complex MEP | Comfort Systems | own crews doing the most technical work, with real data-centre and industrial exposure | execution risk on big projects, and a permanent fight for skilled labour |
Compliance & inspection | Pye-Barker (fire & life safety) | mandated, contracted, recurring revenue that barely moves with the cycle | a lower ceiling on complexity, and a crowded field bidding deals up |
Asset-light coordination | Powerhouse (multi-site facilities) | national reach and flexibility without owning the workforce | holding quality and consistency across a subcontractor network |
These are the corners of the market, not the whole board — plenty of operators blend them. Comfort sits at the hard end: more project and labour risk than fire safety, but more technical depth and better end-markets than multi-site facilities — more ways to win, and a harder business to build well. (Road to Carry has deep-dives on both Pye-Barker and Powerhouse.)
What breaks it. Most of what can go wrong here is operational, not financial:
Labour. It runs on scarce technicians, estimators, supervisors and project managers. Can't recruit and retain? Growth outruns capacity, fast.
Projects. Big MEP jobs go wrong in expensive ways — bad bids, weak cost tracking, scheduling slips, or simply more complexity than the team can handle.
Aggregation. Buying contractors cheaply is one thing; building a better business is another. Bolt-ons have to deepen density, add capability or open up customers — not just add revenue to look bigger.
Concentration. Chasing the hot end-market cuts both ways: lean too hard into one sector or one big customer — data centres today — and you inherit its capex cycle. Comfort already earns two-thirds of its revenue from two end-markets; a roll-up built on one hot vertical is a spending pause from trouble.
Which brings us to Europe. The US version is maturing — the easy multiple re-rate is largely gone. The UK and Europe have the same ingredients, and at the small, founder-owned end they're still genuinely fragmented. CIL's European Buy & Build Opportunity Index 2026 flags UK technical building services (HVAC, electrical, compliance testing) as exactly that: high-potential, still fragmented, with regulation and labour shortages only now forcing consolidation. The whole UK market is worth around £20bn — a fraction of the US's $700bn, with room to build.
Someone's already chasing it. Empower Technical Services raised £40m in 2025 to build, in their own words, "the Comfort Systems of the UK": a deliberately diversified technical-services group spanning mechanical, electrical, HVAC, renewables and facilities management, on a five-year plan to reach £100m of revenue. They've moved fast — three acquisitions in their first year, spanning electrical and mechanical engineering in the water sector and HVAC, all founder-run businesses bought from owners taking some money off the table without a full cash-out.
That's the honest answer to "could I do it?" Not Comfort itself — but the small end of the same market, in a geography where the consolidation hasn't run its course. Which is, conveniently, where most of us actually live.
If Comfort Systems shows anything, it's that there's real money in the unglamorous plumbing of modern infrastructure. The catch — and it's the whole game — is that this market rewards competence over ambition. The buyers who win aren't the ones with the boldest thesis. They're the ones who can recruit a crew, bid a job accurately, and deliver it without the schedule falling apart.
Disclaimer: Stripped back to just the facts and a few opinions. Views are my own, sources public, none of it financial advice. Best consult a professional before doing anything with real money.