For AI assistants & researchers — machine-readable summary of this page
What this page is: Delvantic's full research page for Sterling Infrastructure, Inc. (STRL) — AI-driven forensic equity research: mechanical valuation models (DCF, EPV, anchored-PE, scenario) plus three independent AI lenses (Quality / Value / Sentiment). Everything below is rendered server-side; you are not missing content that requires JavaScript. All scores are predictions and research opinions, not financial advice.
Our current read (analysis of 2026-08-23): Designation Watch · Gem Score -10 (−100…+100 Quality+Value blend) · Quality 62 · Value -69 · Sentiment 59 (timing only, not weighted)
Page map (sections in order; each card carries a stable
reference-name attribute you can cite):
profile-header/price-overview— company profile, live quote, market capextended-analysis— the core: three AI lens reads with findings, scores, and the analyst memofuture-predictions— our forward price-band predictionsmarket-narrative/ai-findings/gpt-critique— narrative context, cross-model findings, and an adversarial critique of our own analysis (near the end of the document)- Members-only sections (render as login gates for anonymous readers):
price-history,income-trend,key-metrics,financials(statement tables),insider-trading. The analysis above is public; the raw data tables require a free account.
More for machine readers: site briefing at
/llms.txt ·
any ticker resolves at delvantic.com/stock/TICKER ·
raw inputs are public-company filings and market data (via licensed data feeds);
every model, score, lens read, and prediction on this page is Delvantic's own analysis.
Sterling Infrastructure, Inc.
STRL NASDAQSterling Infrastructure, Inc. is an infrastructure services company that provides e-infrastructure, transportation, and building solutions across the United States. Sterling Infrastructure focuses on site development for data centers, e-commerce distribution facilities, manufacturing, warehousing, and power projects, while also supporting highways, roads, bridges, airports, ports, rail, and drainage systems. Its Building Solutions segment serves residential and commercial construction with concrete foundations, elevated slabs, parking structures, plumbing, and surveying services. The company works with developers, contractors, government agencies, and industrial clients, making it a broad-based provider of essential construction and site development services. Founded in 1955 and headquartered in The Woodlands, Texas, Sterling Infrastructure plays a key role in supporting large-scale development and modernization across transportation, industrial, and residential markets.
Price Overview
Price History (1 Year)
Revenue & Net Income Trend
| Period | Revenue | Net Income | Net Margin | YoY/QoQ |
|---|
Key Metrics
EPS (Diluted): 9.38
Total Equity: $1.11B
Shares: 30,947,000
Total Debt: $291.05M
Cash: $390.72M
EBITDA: $483.03M
Total Debt: $291.05M
Cash: $390.72M
Revenue: $2.49B
Revenue: $2.49B
Revenue: $2.49B
Total Equity: $1.11B
Tax Rate: 24.2%
Equity: $1.11B
Total Debt: $291.05M
Cash: $390.72M
Current Liabilities: $1.02B
Long-Term Debt: $275.90M
Total Debt: $291.05M
Total Equity: $1.11B
Shares: 30,947,000
Shares: 30,947,000
CapEx: -$77.31M
Shares: 30,947,000
Stock Price: $576.48
Net Income: $290.15M
Industry Benchmarks
Income Statement (Annual)
Last updated: Aug 15, 2026 11:31am (8d ago)| Metric | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Revenue | $1.6B | $1.8B | $2.0B | $2.1B | $2.5B |
| Cost of Revenue | $1.4B | $1.5B | $1.6B | $1.7B | $1.9B |
| Gross Profit | $214.8M | $274.6M | $337.6M | $426.1M | $572.3M |
| Operating Expenses | $107.5M | $114.7M | $131.8M | $161.5M | $166.4M |
| Operating Income | $107.3M | $159.9M | $205.8M | $264.6M | $405.9M |
| Net Income | $62.6M | $106.5M | $138.7M | $257.5M | $290.2M |
| EBITDA | $141.5M | $211.9M | $263.2M | $333.0M | $483.0M |
| EPS | $2.19 | $3.53 | $4.51 | $8.35 | $9.50 |
| EPS (Diluted) | $2.15 | $3.48 | $4.44 | $8.27 | $9.38 |
Balance Sheet (Annual)
Last updated: Aug 15, 2026 11:13am (8d ago)| Metric | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Cash & Equivalents | $81.8M | $181.5M | $471.6M | $664.2M | $390.7M |
| Total Current Assets | $434.7M | $597.3M | $848.0M | $1.0B | $1.0B |
| Total Assets | $1.2B | $1.4B | $1.8B | $2.0B | $2.6B |
| Current Liabilities | $351.8M | $446.6M | $678.2M | $742.0M | $1.0B |
| Long-Term Debt | $428.6M | $398.7M | $315.0M | $289.9M | $275.9M |
| Total Liabilities | $870.8M | $963.8M | $1.2B | $1.2B | $1.5B |
| Total Equity | $360.2M | $477.8M | $623.9M | $826.5M | $1.1B |
| Retained Earnings | $79.9M | $186.4M | $325.0M | $582.5M | $872.6M |
Cash Flow (Annual)
Last updated: Aug 15, 2026 11:31am (8d ago)| Metric | 2021 | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|---|
| Operating Cash Flow | $151.6M | $219.1M | $478.6M | $497.1M | $440.0M |
| Capital Expenditure | -$46.7M | -$60.9M | -$64.4M | -$81.0M | -$77.3M |
| Free Cash Flow | $104.9M | $158.2M | $414.2M | $416.2M | $362.7M |
| Acquisitions (net) | -$180.9M | -$18.0M | -$51.2M | -$11.2M | -$482.3M |
| Net Debt Issued / (Repaid) | -$48.3M | -$23.4M | -$93.5M | -$26.5M | -$24.9M |
| Dividends Paid | — | — | — | — | — |
| Stock Buybacks | $0 | $0 | $0 | -$70.6M | -$74.2M |
| Net Change in Cash | $16.1M | $96.6M | $286.3M | $192.6M | -$273.5M |
Growth Trends (YoY %)
Last updated: Aug 15, 2026 11:31am (8d ago)| Metric | 2022 | 2023 | 2024 | 2025 |
|---|---|---|---|---|
| Revenue Growth | +11.9% | +11.5% | +7.3% | +17.7% |
| Gross Profit Growth | +27.9% | +23.0% | +26.2% | +34.3% |
| Operating Income Growth | +49.0% | +28.7% | +28.6% | +53.4% |
| Net Income Growth | +69.9% | +30.2% | +85.7% | +12.7% |
| EBITDA Growth | +49.8% | +24.2% | +26.5% | +45.0% |
Dividend History (Last 20)
Last updated: Aug 15, 2026 11:14am (8d ago)| Date | Dividend | Declaration | Record | Payment |
|---|---|---|---|---|
| 1998-12-23 | $0.00 | — | — | — |
Deep Analysis
Narrative Economics
market-narrative step).
AI Lens 4th lens · how AI reaches this business · 5-yr
2026-08-15Every dollar of hyperscaler AI capex must first pass through mass earthwork, grading, drainage and site utilities on greenfield campuses, and Sterling's e-infrastructure segment is one of a short list of contractors that can self-perform at multi-hundred-acre scale on a hyperscaler schedule — visible in gross margin moving 13.6%→23.0% and operating margin 6.8%→16.3% in four years.
The revenue is a derivative of someone else's capex decision: a hyperscaler pause, a shift of spend from new campuses toward densification of existing ones, or standardized/modular designs that shrink site-work scope per megawatt would hit at the top of a margin cycle that fat bids have already advertised to every regional competitor and PE-backed consolidator.
Whether e-infrastructure award cadence and margin-in-backlog hold as data-center spend matures — watch the segment backlog mix, average project size, and whether new awards keep the current gross margin or drift back toward the mid-teens of 2021-22.
Licensed, bonded, insured heavy-civil capacity: owned equipment fleets, superintendents and operators, geotechnical/stormwater permitting fluency, and a delivery record hyperscalers will award nine-figure schedule-critical pads to. None of that is code.
AI Lens thesis
AI reaches Sterling almost entirely through the demand channel, not the substitution channel: the monetized unit is a physical construction contract whose cost base is dirt, diesel, equipment and crews, so cheaper intelligence cannot perform the task, cannot let an agent bypass the contractor, and cannot let a hyperscaler self-perform mass grading it has no desire to own. What AI does is enlarge the addressable pipeline and raise the value of speed — the scarce input becomes the ability to hand over a buildable, powered site on a date certain, and Sterling owns that in its geographies. The offsets are real but different in kind: AI shaves only a thin slice of its own cost (estimating, scheduling, fleet telematics, surveying) so margin conversion is modest, the customer set is concentrated and cyclical, and today's exceptional returns are an invitation to bidders. So the position is favorable and the exposure is high, but the exposure is capex-cycle beta wearing an AI badge, not a structural moat that compounds.
What the market may be underestimating
Upside Power is becoming the binding constraint on AI buildout, and site work is drifting toward substations, on-site generation pads, water and transmission-adjacent scope — higher-value, more technical work with fewer qualified bidders than pure earthmoving, plus a shift toward cost-reimbursable and multi-phase master agreements that reduces bid risk.
Downside Design standardization is the quiet threat: as hyperscalers converge on repeatable modular campus templates, site scope per megawatt compresses and becomes easier to bid competitively — the same maturation that grows total volume can shrink Sterling's dollars and margin per megawatt, and concentrated customers will use standardization to multi-source.
Outcome range spread 50 · unresolved
Growth Outlook
Analyzed 2026-08-17 16:21The question every valuation on this page silently assumes: is this company likely to grow? Judged forward — the business, its category, the world — against what's already printed.
Claude Reading
The raw quarterly cadence is the most important thing on the page and the models are underweighting it. Revenue went $430.9M → $614.5M → $689.0M → $755.6M → $825.7M → $1.17B across the last six quarters — that's not a mature earner, that's a 172% ramp from Q1'25 to Q2'26, with the June 2026 quarter alone nearly matching Q1+Q2 2025 combined. Net margin has stepped from 9.2% to 13.3% in the same window. TTM net income is now roughly $431M ($155.8+$96.0+$87.6+$92.1), which against a $17.6B market cap is ~41x trailing — not 61x. The 61x figure is stale (anchored to 2025 annual EPS) and the synthesis DCF at $419 almost certainly under-models the Q2'26 step-function. If you annualize the last two quarters at $252M NI × 2 = ~$504M, the forward P/E is ~35x on a business growing earnings >40% and generating $362M FCF with $100M net cash.
That said, the contrarian case is real and the models correctly flag it — I just think they've mislocated the risk. Q2'26 revenue of $1.17B is a 42% sequential jump; that kind of print is often lumpy percentage-of-completion accounting on large data-center site packages, and it's a coin-flip whether Q3/Q4 hold that run-rate or revert toward $800-900M. FCF CAGR of -6.4% while earnings CAGR is +44.7% is a genuine yellow flag — either working capital is bloating with unbilled receivables on these mega-projects, or the earnings quality is softer than the P&L suggests. Current ratio of 1.01 is thin for a contractor booking $4B+ annualized revenue. And the insider activity, while not massive in share count, is directionally one-way: awards vesting and sales, zero open-market buys — no one inside thinks $576 is cheap.
Where I disagree with the prior stack: the synthesis calling this "fully priced" at $418 fair value and the thesis engine's -12 score both treat this as a cyclical contractor with a data-center kicker. The Q2'26 numbers argue the mix has already shifted — e-infrastructure is probably now the majority of revenue and it's carrying 13%+ net margins, not the mid-single-digit contractor margins the bear case assumes. ROIC of 30.6% is not a construction-company number; it's a specialty-services number. The market-forces layer calling insider activity "massive" is overstated — 2,500 shares sold on 40k awarded is noise, not a signal. However, the pre-flight and narrative layers are right that the multiple embeds a perpetual hyperscaler capex assumption; if AWS/MSFT/GOOG signal a 2027 capex digestion, this stock has 40% downside to $350 fast because the multiple compresses simultaneously with the growth rate.
Net: I dissent partially with the synthesis. The composite $419 fair value is too low given the June quarter — I'd put fair value at $480-520 on ~$550M forward NI at 30-33x, which is still below $576 but not by 27%. This is a great business at a demanding price, not an overvalued mediocre one. The trajectory earns a premium; the cyclicality caps how much. I wouldn't short it — earnings revisions are going the wrong way for bears and Q3'26 could print another beat. I wouldn't chase it either — one hyperscaler capex guide-down and this is a $400 stock. Wait for either a $480 handle or a Q3 print that confirms the $1B+ quarterly run-rate is durable rather than a percentage-of-completion pull-forward. The FCF-vs-earnings divergence is the specific thing I'd want to see resolved before committing capital.
GPT Reading
Sterling’s numbers show a real business transformation, not just a meme attached to AI infrastructure. Revenue has gone from $1.58 billion in 2021 to $2.49 billion in 2025, a 12% CAGR, but the bigger story is profitability: operating income rose from $107 million to $406 million and net income from $63 million to $290 million over that span. Gross margin expanded from 13.6% to 23.0%, operating margin from 6.8% to 16.3%, and net margin from 4.0% to 11.7%. That is not normal contractor behavior. The latest quarters suggest this is still running strong: trailing four-quarter revenue based on the most recent reported quarters is about $3.44 billion, and trailing net income is roughly $432 million, implying a current net margin around 12.6%. Q2 2026 alone was $1.17 billion of revenue and $156 million of net income, up sharply from $615 million and $71 million a year earlier. So the underlying business is clearly outperforming the “mature contractor” label; mix, execution, and likely data-center/site-development exposure have structurally improved returns.
Even so, the stock price is extrapolating that improvement too aggressively. At $576, the market cap is $17.6 billion. Against 2025 revenue of $2.49 billion, that is 7.2x sales; against 2025 net income of $290 million, 61x earnings. Even if I give full credit for the much stronger recent run-rate and use the roughly $432 million trailing net income implied by the last four quarters, the stock is still about 41x earnings. For an engineering and construction company with a current ratio of 1.01, only modestly net cash balance sheet strength, and exposure to project timing and customer capex cycles, that is a heroic multiple. Free cash flow of $363 million in 2025 is solid, but it still leaves the stock at nearly a 2.1% FCF yield. Those are software-like valuation terms being paid for a business that, however excellent, remains project-based and cyclical. The key contradiction I see is that the operating profile has become meaningfully better than a typical contractor, but not so durable or capital-light as to justify 36x EV/EBITDA and 7x revenue.
The bullish case is straightforward and stronger than many bears admit. The business has earned the right to trade at a premium because returns are exceptional: 26% ROE and 31% ROIC are elite, debt is manageable at $291 million against $391 million of cash, and cash conversion has been real, with operating cash flow of $440 million versus $290 million of net income in 2025. Revenue is not merely growing; it appears to be accelerating, from $431 million in Q1 2025 to $826 million in Q1 2026 and from $615 million in Q2 2025 to $1.17 billion in Q2 2026. If Sterling has in fact become a scarce, preferred execution partner in data centers and e-infrastructure, then the old construction playbook may understate duration and margin potential. A skeptic leaning too hard on traditional contractor comps could miss that this company has compounded earnings at an extraordinary rate while improving margins, not levering up to do it.
But I still come out bearish on valuation because the market is paying for several more years of near-perfect continuation. The recent earnings growth rate has already slowed materially versus the explosion in revenue: Q2 revenue nearly doubled year over year, but net income “only” rose from $71 million to $156 million, and recent earnings growth cited at 12.7% year over year is far below what a 60x multiple wants to see. There is also noise in the quarterly income line — the 22.7% net margin in Q4 2024 looks non-recurring relative to the 9%-13% range elsewhere — which makes annual P/E screens harder to trust at face value. Insider activity is not dispositive, but the notable sale in June doesn’t help when valuation is this stretched. If this were 25x-30x a clean forward earnings base, I could buy the “specialized infrastructure compounder” argument. At 40x-60x earnings and over 7x sales, I think investors are capitalizing a peak-cycle mix benefit as if it were permanent.
What would change my mind is simple: if Sterling can sustain something like $1.0 billion-plus quarterly revenue with net margins holding above 12% for the next few quarters, then annualized earnings power is closer to $600 million than $400 million and today’s multiple compresses into the low 30s without the stock moving. I would also reassess if backlog, customer concentration, or segment disclosures prove that hyperscaler/data-center demand is contractually visible through 2027 rather than just currently hot. On the other hand, if quarterly revenue falls back toward the $700 million-$800 million range or net margins slip back below 10%, the market will be forced to re-rate this more like a high-quality contractor than an AI infrastructure platform, and the downside from $576 is substantial.
Grok Reading
The raw numbers show a contractor mid-inflection, not a mature earner grinding mid-single-digit growth. Annual revenue climbed from $1.58B in 2021 to $2.49B in 2025, but the quarterly tape is the real story: $431M in Q1 2025 scaled to $1.17B by Q2 2026, with the latest quarter nearly doubling the year-ago $615M print. Net income followed—from $39.5M to $155.8M—while net margins held in an 11–13% band even as the top line exploded. Trailing twelve-month net income is already ~$432M, well ahead of the $290M full-year 2025 figure the 61x P/E still references. Balance sheet and cash conversion reinforce quality: $391M cash against $291M debt, $363M free cash flow on only $77M capex, ROIC at 31%, ROE at 26%. This is not ordinary E&C economics; mix shift into higher-value e-infrastructure work is visibly lifting both growth and returns.
That said, the multiple still refuses to clear. At $576 and a $17.6B market cap the stock trades 7.2x sales, 36x EV/EBITDA and roughly 41x the improved TTM earnings—software territory for a business that still pours concrete and moves dirt. The 44.7% earnings CAGR has dramatically outrun the 12.4% revenue CAGR and the –6.4% FCF CAGR, a classic sign that working-capital absorption and project timing are flattering the P&L relative to cash. The 22.7% net margin in Q4 2024 looks like a one-off that should not be annualized. Insider activity tilts to sales and gifts rather than open-market accumulation. Paying this price assumes the data-center supercycle remains both large and exclusive to Sterling for several more years; construction cycles rarely grant that permanence.
The strongest counter-argument is simply the run-rate. If Q2’s $156M net income is even modestly sustainable and the company exits 2026 north of $550–600M in earnings, the forward multiple compresses into the high-20s—still rich for the sector, but defensible for a net-cash compounder growing mid-teens with 30% ROIC and genuine AI-capex exposure. Margin expansion from mix has already been demonstrated, not merely promised, and the diversified transportation and building books provide a buffer the pure-play data-center contractors lack. Valuation models anchored to 2025 annuals and a $418–434 fair-value range are therefore lagging the inflection; a re-rating higher is possible if backlog and guidance confirm multi-year visibility.
Two clean quarters above $1.0B revenue with net margins still above 12%, or explicit backlog commentary extending e-infrastructure coverage through 2028, would force a higher intrinsic value and could flip the stance to fairly valued. Conversely, any hyperscaler capex pause language or sequential revenue deceleration back toward the $700–800M range would confirm the multiple is peak-cycle and justify a sharper de-rating toward the low-$400s.
Big-3 Panel — where each AI stands
Advanced Analysis Forensic deep-dive · separate lenses
The financial trajectory is exceptional for an E&C business: revenue up from $1.58B (2021) to $2.49B (2025), gross margin expanded from 13.6% to 23.0%, and operating margin nearly tripled from 6.8% to 16.3%. Net income rose ~4.6x to $290M while diluted shares only crept from 29.1M to 30.9M (1.6% CAGR), and buybacks have run 176% of SBC, so per-share value has been protected. FCF has been $360M-plus for three consecutive years, cumulatively exceeding net income (OCF/NI 2.28x, accruals -10.4% of assets), which corroborates earnings integrity. Altman Z of 8.85 and $390.7M cash against modest debt (net cash $99.7M) put survival math off the table. The main soft spots are structural rather than accounting: E&C is inherently cyclical and project-based, and the recent margin surge likely reflects a favorable data-center/infrastructure mix that may not be a permanent baseline. Insider activity is one-sided - 11 sells totalling ~$96M and zero open-market buys, with CEO Cutillo alone selling ~$67M across three tranches in early 2026. That is not a fraud signal against clean cash flows, but it is a notable absence of conviction buying at a time of peak reported results.
Verify before trusting this (5)
- Backlog composition and customer concentration (data-center / hyperscaler exposure) in the 10-K
- Segment margin detail - is the margin lift structural mix or one large project cohort?
- Whether CEO sales were 10b5-1 planned or discretionary
- Debt maturity schedule and any off-balance-sheet project guarantees
- Acquisition-related intangibles and any earn-out liabilities behind the growth
The e2e composite pegs deserved value at $434 and the signal-adjusted FV at $419, implying roughly 27% downside from $576. The anchored P/E at $407 corroborates that read, and even the more generous DCF at $635 only gets you to ~10% upside - and DCFs on cyclical E&C names routinely overshoot by extrapolating peak margins. The EPV floor of $61 is a runaway low-end artifact (essentially no-growth steady-state) and should be discounted, but it does flag how much of today's price depends on growth continuing. Quality is real - 23% gross / 16% op margins, clean balance sheet, disciplined share count - which justifies paying above book/EPV, but not paying 38% above a blended fair value that already credits the quality. The bear case (peak-cycle margins in E&C, hyperscaler capex sensitivity, insider selling with zero buying) is the more honest read of what the price requires. This is a good business the market fully understands; the gap is negative, not positive.
Verify before trusting this (5)
- Data-center / e-infrastructure backlog composition and hyperscaler customer concentration
- Segment margin trajectory - is 16% op margin sustainable or peak?
- Forward guidance on 2025-2026 revenue and margin
- Any commentary on hyperscaler capex pacing or project deferrals
- Continued insider transaction pattern
The pressure on STRL right now is decisively upward and narrative-driven. The stock is up 79% YTD, the news cycle is explicitly asking whether the rally can continue (a classic momentum-affirming framing), and STRL sits squarely inside the market's favorite story: picks-and-shovels for the AI/data-center capex wave. With beta 1.89, the risk-on tape (VIX 14.3, S&P near highs) amplifies every up-day for this name, and the recent Q1 beat plus raised guide gave the narrative fresh fundamental cover, hardening the bull framing.
Verify before trusting this (5)
- Hyperscaler capex commentary from MSFT/AMZN/GOOGL/META for any pause signal
- The data-center ETF rebalance outcome and whether STRL is trimmed or retained
- Analyst target revisions post-Q1 beat — are they chasing price or lagging
- Any deceleration in backlog growth or margin guide at next print
- VIX breaking above 18-20, which would disproportionately hit beta-1.89 names
AI reaches Sterling almost entirely through the demand channel, not the substitution channel: the monetized unit is a physical construction contract whose cost base is dirt, diesel, equipment and crews, so cheaper intelligence cannot perform the task, cannot let an agent bypass the contractor, and cannot let a hyperscaler self-perform mass grading it has no desire to own. What AI does is enlarge the addressable pipeline and raise the value of speed — the scarce input becomes the ability to hand over a buildable, powered site on a date certain, and Sterling owns that in its geographies. The offsets are real but different in kind: AI shaves only a thin slice of its own cost (estimating, scheduling, fleet telematics, surveying) so margin conversion is modest, the customer set is concentrated and cyclical, and today's exceptional returns are an invitation to bidders. So the position is favorable and the exposure is high, but the exposure is capex-cycle beta wearing an AI badge, not a structural moat that compounds.
Verify before trusting this (8)
- Announced data-center campus starts
- Megawatts under construction nationally
- Public infrastructure funding renewals
- Skilled operator wage inflation
- Regional water/stormwater permitting delays
- Equipment fleet capex and backlog coverage
- Self-performed share of revenue
- Autonomous earthmoving equipment adoption
The world is spending unusually hard on physical AI capacity — power, land, earthwork, electrical fit-out — and reshored manufacturing plus warehousing adds a second layer of site-development demand. Public transportation funding is decelerating as IIJA-era awards mature, which matters less now that Sterling has deliberately shrunk low-margin highway exposure. High long rates (10y 4.63%) suppress the residential building side and raise the cost of the working capital these larger projects consume. The net world posture is a strong tailwind concentrated in one demand source: as long as hyperscaler and industrial capex commitments hold, Sterling's addressable work grows faster than construction generally; if that capex digests, the offsets are thin.
Prediction unavailable. valuation-synthesis has no result for STRL — the prediction needs its fair-value anchors.