Everyone's been chasing the AI bottleneck down the stack, GPUs, then power, then memory, then optical. The one trade that quietly ranked top-three every single year is the boring one nobody posts about: data center construction. Sterling Infrastructure ($STRL) is the closest thing to a pure-play on the part of it that actually has a moat, site work. That means clearing, grading and blasting the land before a building exists. It sounds like dirt, and it is, but the economics are good because the business is regionally locked (you can't truck a fleet of bulldozers across the country and stay cost-competitive) and hyperscalers won't split a 1,000-acre pad across two contractors when a schedule slip costs them tens of millions a day. Sterling is ranked #1 in site work, sits right on top of the Virginia, Texas and Georgia clusters, and just bought CEC to bolt electrical and mechanical work onto the site prep so it can run both in parallel. 1Q26 was the tell: revenue up 92%, adjusted EPS of $3.59 against a $2.29 estimate, backlog up 55%.
Here's my hesitation, and it's the whole debate. The valuation already prices the dream. The bull case I read puts a 42.5x forward multiple on 2027 EPS to get a $1,500 target, basically saying "treat it like Quanta." That's not a value setup, that's paying up for growth that has to show up exactly on schedule, and a chunk of the thesis leans on cross-sell synergies that are one or two quarters old. The business quality looks real to me (self-performance, PM bench, 14%+ FCF margin), but at 40-plus times earnings you're underwriting flawless execution and a DC capex cycle that doesn't blink. I'd rather watch it and buy the first time the AI capex narrative wobbles and this thing gets cut in half, because a name this cyclical will give you that chance. Position: no position, on the watchlist. For anyone who owns it here, what's your margin of safety at this multiple, or are you purely riding the backlog?