Business
Know the Business
Sterling is no longer the heavy-highway contractor it was a decade ago. It is now, effectively, a mid-cap data-center site-development specialist with two adjacent revenue streams (DOT roads and Texas residential foundations) bolted on. The market is correctly excited about the data-center engine and the new electrical-services arm (CEC, acquired Sept 2025) — what it tends to under-appreciate is how narrow the moat is (project-management bandwidth and customer relationships, not technology) and how uncyclical the company has chosen to be by deliberately ceding low-bid work where margins collapse first.
Revenue (FY25, $M)
▲ 17.7% YoY (ex-RHB)
Operating Margin
Backlog YE25 ($M)
Return on Equity
1. How This Business Actually Works
Sterling is a project-execution shop, not a product company. It sells time-and-materials risk transfer to customers who cannot afford a missed schedule — hyperscalers building data centers, state DOTs running a federally funded paving program, homebuilders who need 200 foundations poured this month. The economic engine is project selection × project execution × bonding capacity, in that order.
The segment math is the whole story: E-Infrastructure is 59% of revenue but 75% of segment operating income because it earns roughly 24% margins versus 12% in Transportation and 10% in Building. Within E-Infrastructure, ~80% of backlog is "mission-critical" work (data centers, semiconductor fabs, e-commerce distribution), and within that, data-center revenue grew 125%+ year-over-year in Q3 FY2025. So when you buy STRL, ~70%+ of incremental dollars come from one customer cohort — hyperscalers and colocation providers building AI/cloud capacity.
Where the bargaining power sits. Sterling sits in the middle of the value chain: the customer (Amazon, Meta, Google, a colocation REIT) decides what to build and where; the GC or developer hires Sterling for the dirt and pad work; subcontractors and equipment dealers supply Sterling. Pricing power on any one project is modest — these are competitively bid — but Sterling's ability to finish on schedule lets it skip past the bottom of the bid stack. Management has explicitly said it will pass on megaprojects if pricing or contract terms are not right, which only works because backlog already covers ~14 months of revenue.
What turns into incremental profit. Two things compound: (i) mix shift toward larger projects — bigger data-center sites get the same overhead spread over more square feet, and (ii) integration of CEC — bundling site development with electrical/mechanical work historically lifted a small dry-conduit tuck-in's margins by 40%, and management believes CEC will see similar lift over a couple of years. Conversely, what compresses margins fastest is a single bad project estimate — the cost-to-cost revenue recognition method recognizes losses immediately, which is what drove the FY2011–FY2016 wreck.
Mental model: picture Sterling as a fleet of 100–150 active projects with a portfolio bid margin around 17–18%. Earnings power is roughly: backlog × execution margin – G&A. The backlog tells you nine months of revenue and the bid margin in backlog tells you the next year's gross margin floor.
2. The Playing Field
Sterling sits in an awkward but profitable spot in the U.S. infrastructure-services peer set: smaller than the megaprime PWR/MTZ, larger and pickier than the regional pure-plays, and uniquely concentrated in data-center site work rather than utility transmission or telecom fiber.
What this peer set actually says.
Sterling is the highest operating-margin name in the entire U.S. infrastructure-services public peer set — by a wide margin. PWR and MTZ are 3–4× larger but earn 5–6% operating margins because utility-T&D work is a thin-margin volume game; STRL earns ~15% because it has chosen to specialize in time-sensitive site development where the customer cares more about certainty than price. The "good in this industry" benchmark is no longer 6% op margins — Sterling has redefined it.
The closest analog is DY (Dycom) — a mid-cap specialty contractor that found its niche in telecom/fiber and now grows alongside one customer cohort (broadband providers) the way Sterling grows alongside hyperscalers. DY trades at a similar EV/EBITDA (~19×) but carries 3.3× net leverage versus Sterling's net cash. ROAD is the cautionary peer — fast revenue growth (+54% YoY) bought largely through M&A and 4× net leverage, with margins still well below Sterling's despite a similar specialty story.
Sterling's weaknesses relative to the set are subtler. It has the smallest revenue base of the group (~$2.5B vs PWR at ~$28B), which limits how many simultaneous megaprojects it can absorb without project-management strain. Management has acknowledged this: "if somebody comes in tomorrow and says 'I've got three $500M jobs that need to start in 60 days,' that would be a challenge." That's the constraint that determines whether the next leg of growth is execution-led or stall-out.
3. Is This Business Cyclical?
Yes, but the company has engineered out the worst of the cyclicality since 2016 by exiting the work that bleeds most in a downturn. The historical record makes the change visible.
Where the cycle hits, in this business specifically:
| Lever | What happens in a downturn | How exposed is Sterling now? |
|---|---|---|
| Demand volume | DOT bids slow; private capex deferred | Low — IIJA funded through Sept 2026, hyperscaler capex multi-year |
| Project pricing | Bidders cut margin to keep crews busy | Medium — convergence into mid-market is the named risk in the 10-K |
| Cost-to-cost re-estimates | Overruns on fixed-price work hit the P&L immediately | Medium — most backlog is fixed-unit-price or lump-sum |
| Working capital (Contract Capital) | Receivables stretch, bonding capacity tightens | Low — net cash, $390M cash, undrawn revolver |
| Residential exposure | Building Solutions tracks new home starts | Real and live — Building op income fell from $54M to $39M FY24→FY25 |
The FY2011–FY2016 hole is instructive. Revenue actually grew through that period (from $501M to $690M) — Sterling did not have a demand problem. It had a bid-discipline problem: low-bid heavy-highway projects with ~4% gross margins gave no cushion when costs ran. A handful of large under-bid projects produced a $69M operating loss in FY2013 alone. The 2016 strategic pivot — shrink low-bid heavy highway from 79% of revenue down to 9%, push into alternative-delivery DOT work, then build E-Infrastructure into the leading segment — is exactly the lesson that produces today's 14.8% margin. Whether that lesson holds up under a real data-center pause is the open question.
The visible cyclical pressure right now is in Building Solutions, where residential-foundation revenue declined as homebuyers struggle with affordability. Management has guided to a mid-to-high single-digit revenue decline there, with margins compressing from 14.8% to low double digits. This is a useful tell: when Sterling is hit by a real housing downturn (rates, affordability), the segment loses ~3 points of operating margin and ~25% of segment operating income, but the consolidated business absorbs it because E-Infrastructure is structurally larger.
4. The Metrics That Actually Matter
Forget the standard P/E and dividend-yield checklist for this business. Five metrics tell you almost everything about whether earnings are about to expand, hold, or compress.
The backlog tells the cleanest version of the story: it nearly doubled in twelve months (the CEC acquisition contributed ~$489M, but the rest was organic — book-to-burn excluding CEC was 1.31x year-to-date through Q3). Crucially, margin in backlog rose 110 basis points at the same time. That combination — more dollars at higher margins — is the single most reliable signal that the next year's earnings power is structurally higher, not just bigger.
Two further notes for an analyst. First, Contract Capital (receivables + contract assets – payables) has been a $54M use of cash in FY25 versus a $186M source in FY24, driven by E-Infrastructure project size and duration. A growing business funds working capital — that's normal and healthy here, but watch the ratio of operating cash flow to net income. FY25 came in at ~1.5x; if it slips below 1.0x for a full year, ask why. Second, customer concentration in E-Infrastructure (top 4 = 27% of segment revenue, down from 40% in 2023) is improving but still a real exposure. A single hyperscaler pulling capex would not be a rounding error.
5. What I'd Tell a Young Analyst
This is a specialty execution business priced like a growth stock because the data-center cohort is, for now, growing like one. Three things to internalize:
Watch backlog and margin-in-backlog every quarter, not the EPS print. Construction P&Ls are noisy because cost-to-cost revenue recognition pulls future profit forward and pushes losses back. The metric the company itself uses internally — book-to-burn ratio — is a far better leading indicator than the headline beat. If book-to-burn drops below 1.0x for two consecutive quarters, the multiple compresses before earnings turn.
The market is probably right about data centers and probably wrong about the cycle floor. The bull case (multi-year hyperscaler capex, $4B+ E-Infra pipeline, IIJA reauthorization on track) is plausible and largely in the price. What is not in the price is what happens to the multiple if E-Infrastructure organic growth steps down from 30%+ to mid-teens — Sterling has trained the buy-side to expect 25%+ E-Infra revenue growth indefinitely, and that math breaks at some point. The realistic risk is not a 2008-style crash; it is a deceleration that leaves earnings flat for a year and the multiple halved.
The CEC integration is the most underwritten variable. It is a $562M deal (~6% of market cap), it is the largest acquisition Sterling has done in a decade, and the entire investment thesis on bundled site-plus-electrical pricing power has not yet been tested at scale. The earlier dry-conduit tuck-in delivered 40% margin uplift, but that does not generalize to a $300M+ revenue electrical business. Track CEC's standalone operating margin disclosure each quarter — if it is not visibly climbing by mid-2026, the cross-sell thesis is weaker than presented.
What would change the thesis (in either direction):
Bull breaks higher: CEC margins lift 200 bps in 12 months; STRL wins three megaprojects in Texas; IIJA reauthorization passes early.
Bear case activates: A hyperscaler announces a capex pause; book-to-burn drops below 1.0x for two quarters; a single large data-center project takes a material cost-overrun charge; CEC integration produces a goodwill impairment.
The right way to think about valuation is not "what's the right P/E for E&C," because E&C P/Es bracket a 4× range depending on cycle position. The right anchor is EV/EBITDA at a normalized 12% E-Infra growth rate — assume the data-center super-cycle eventually mean-reverts and ask what the business is worth then. If you can defend the price at that normalized state, you have a position. If you cannot, you are paying for a momentum extension that the company itself has not promised.