Sterling Fell Half Its Value While the Business Accelerated: The Sell-Off Fixed One Problem, But It Didn't Make the Stock Cheap

Sterling Infrastructure (STRL) just lost about half its value — from a peak just above $1,000 in June to roughly $511 by late August. In that same stretch the company posted a record quarter: revenue up 90% in the period, a clear earnings beat, and a raised full-year forecast. The chart and the business are telling two different stories, and the gap between them is the whole story. The question worth your time is which one to trust — and what that gap does to the price.

The trade chatter after the drop sounds technical: the rally was a "bull trap," or maybe a "bottom" is forming — and with the stock up about 11% over the last five sessions (and another ~5% on the day, to roughly $511), that is a fair read of the past two weeks. It is also the wrong place to start. Price action is the timing layer. Whether the price is now earned is a fundamentals question, and that is where the judgment actually lives.

The mix the market misread

Sterling started as a civil contractor — it graded the land and poured the foundations. Over the last year it has bought its way into the juicier electrical work inside the data centers, most recently through CEC Facilities Group. Electrical work is far less profitable per dollar than the high-end site development the company is known for. In the most recent quarter the electrical business grew 140%, but it earns about an 11% margin, against the upper-20s margin of the legacy site work.

So when you bolt a fast-growing, low-margin line onto a company, its overall margin falls — even if every single line of the business improves. It is like a restaurant adding a cheaper, higher-volume menu: sales jump, but profit per dish drops. Management said as much: the rapid growth of the lower-margin electrical business diluted the blended margin. And the numbers agree — E-Infrastructure's operating margin came in around 24% for the quarter, down from the high-20s a year earlier. Every component got better; the blend only looked weaker because the mix shifted toward the cheaper work. The market, which had just watched the stock more than double, read that drop as the growth getting old. It was dilution the company chose, not decay.

The selloff fixed one problem

The de-risking was real. At the June peak, Sterling traded at more than 50 times this year's expected earnings. At $511 it is about 25 times — the price divided by the earnings the company expects to make this year. The multiple has essentially halved, which is the market saying, in the only language it trusts, "I do not believe the froth." That is a healthy correction, not a bug.

But "no longer frothy" is not "cheap," and this is the part the technical-bottoming crowd skips. Two things keep the multiple from becoming easy. The first is that growth is slowing — still fast, but visibly stepping down: revenue grew 90% in the most recent quarter, the full year is guided up about 64%, and next year is expected to grow closer to 25%, with less after. The market prices the speed of growth, so a business that is still accelerating in dollars can still get de-rated as its rate of acceleration steps down. The second is that at roughly 25 times forward earnings Sterling still trades at a premium to the broad market. It is not sitting at a near-market multiple where the burden of proof has flipped to the bears. It has left the most dangerous zone. It has not become a bargain.

The one number the price now assumes

So the question is not "did the stock fall too much?" It is whether one specific bet pays off, because that bet is now what the price is paying for. The entire margin story turns on the electrical business Sterling just bought. Management guides it from around 11% margins toward 20% over three to five years, with 300 to 500 basis points of improvement expected in the next 12 to 18 months. If CEC actually re-rates, the blended margin climbs back up and the "decay" narrative dies. If it stalls, the ceiling on the multiple comes down with it. That is the number to watch — and it will show up over the next two quarters, not in this week's candle.

The second dependency is concentration. Ninety-two percent of Sterling's E-Infrastructure backlog is mission-critical — data centers, semiconductor plants, manufacturing. Sterling is, in effect, a pure bet that the AI data-center buildout keeps being funded into 2027 and 2028. That backlog is now $5.6 billion, up 150% year over year, with a total visible pool of work above $7 billion — the bull case and the risk in the same sentence. If those 2025–26 builds are not matched by real utilization later, growth steps down faster than the model assumes and 25 times becomes hard to defend.

What the drop did not break is worth stating plainly. The company has $464.5 million in cash on hand to fund the acquisitions that are doing the growing. Returns on capital are in the 30s. And management guided up the year to about $4 billion in revenue and $20 in earnings per share after a quarter where revenue nearly doubled. This is not a broken company wearing a cheap suit. It is a strong company still at a price that assumes the AI buildout keeps funding itself.

One caution the bulls sometimes wave away: the people running the company have been selling. Insiders logged roughly $47.5 million of sales by the CEO alone over the past six months. For the executives of a stock that had nearly quadrupled in a year, that is position management, not a warning shot — the CEO raised the guidance, after all. But it is a fact, and it belongs on the short side of the ledger.

So here is the posture, and I am being plain rather than dressing it up as a rating. The setup is more honest now than it was at $1,000: the multiple did the de-risking work, and the "selloff" is largely a mix shift the company made on purpose. But at about 25 times forward earnings, with growth decelerating and the electrical-margin bet still unproven, this is neither a value setup nor a chase. The bounce you are watching is the timing layer, and it can time an entry — it cannot make the thesis. The lower-risk move for a new buyer is to let the next two quarters of CEC margin evidence, or a deeper pullback, do the work, and to know exactly what would change your mind: an electrical margin that refuses to climb, or a data-center buildout that never gets repaid. Watch the margin, not the candle.