Every cycle has a company that looks like it will grow forever, right up until the moment it doesn't. Comfort Systems USA might be the current candidate for that lesson, though I'll grant the ride has been genuinely impressive. A recent deep dive on Seeking Alpha lays out the case: the mechanical and electrical contractor is enjoying a surge in revenue and margins thanks to data center construction, and 2026 may mark the peak growth rate before things flatten out.

Start with what's going right, because it's real. Comfort Systems installs the heating, cooling, and electrical infrastructure that data centers need, and AI build-outs have turned that work from steady to frantic. The margin surge is what catches my eye. Contractors, in my long experience following them, run on margins that are modest and hard-won, so when a report describes a significant surge, the pricing environment has genuinely shifted. When your customers are racing each other to pour concrete for server halls, you get to name your price, to a degree.

Now the uncomfortable part, and I've watched this movie enough times to know how it ends. Contractors are cyclical businesses in a growth-stock costume, in my view. Revenue in this kind of work is project-based, so it arrives in lumps, and my concern is that eventually the lumps get smaller. The Seeking Alpha analysis suggests 2026 is when the growth rate crests. Not when revenue collapses, mind you. Just when the year-over-year comparisons stop being spectacular and start being, well, normal.

Some simple math on why peak growth matters so much for the stock. If a company grows earnings 30 percent and the market prices it for 30 percent growth, the stock can hold its valuation while doing nothing surprising. When growth decelerates to 12 percent, though, the multiple usually compresses to match, and the stock can fall 30 percent without the business getting one bit worse. The company didn't fail. Expectations did. I saw this with housing-related names in 2006 and with fiber optics contractors after the dot-com build-out, when perfectly competent companies had their stocks cut in half because the build-out simply finished.

The bear argument writes itself. Data center construction is driven by a handful of enormous spenders making bets on AI demand that hasn't fully materialized yet. If even two of them slow their capital plans, today's pipeline refills slowly. And electrical and mechanical contracting is not a business with deep moats; competitors see fat margins and bid aggressively to get a piece of them. Margins this good in a contracting business attract company.

The bull argument has substance too. The AI infrastructure build-out could run years longer than skeptics expect, and the specialized labor Comfort Systems brings to that work is not easy to replicate overnight. A peak in the growth rate is not a decline in revenue. A company can grow 10 percent a year for a decade after the fireworks end and reward patient shareholders handsomely.

But that outcome depends on what you pay. A stock priced for perpetual 30 percent growth and one priced for a normal cyclical contractor are two very different investments wearing the same ticker. If 2026 is the crest, the question isn't whether the business is good. It's whether today's price already assumes the good times never stop.

I've owned cyclical stocks before, and the ones that treated me best were bought when the news was bad, not when margins were surging and everyone was drawing straight lines into the future. The straight lines are the tell. They were drawn for housing in 2005, for fiber in 1999, and they get redrawn every cycle with fresh enthusiasm.

If you own it, ask yourself whether you'd still want it if revenue growth halved next year. If you don't own it, ask whether you're buying a business or buying momentum with extra steps. Two very different purchases, those.

The usual disclaimer, as always: I'm sharing perspective from four decades of watching cycles roll through, not professional financial advice. Do your own math, and be suspicious of anything that looks like a straight line.