Sterling Infrastructure

Phase A — Understand the business

Lens 1 · Company Overview

Sterling Infrastructure, Inc. (NASDAQ: STRL; HQ The Woodlands, TX; CIK 0000874238) is a U.S. infrastructure construction-services holding company operating through three reportable segments:

• E-Infrastructure Solutions — "advanced, large-scale site development services and mission-critical electrical services for data centers, semiconductor fabrication, manufacturing, distribution centers, warehousing" and power generation. The growth engine. FY2025 revenue $1,466.8M, 59% of total. Q1-2026 it was 72% of revenue.
• Transportation Solutions — heavy highway, aviation, rail; customers are state DOTs and transit/airport/port/rail authorities. FY2025 $640.7M, 26%.
• Building Solutions — residential concrete slab + plumbing (Texas/Phoenix) and commercial. FY2025 $382.6M, 15%.

Business model. A project-based EPC/specialty-contractor model: Sterling bids on and executes construction contracts (lump-sum, fixed-unit-price, and residential/other). Q1-2026 contract mix: Lump-Sum $533.9M / Fixed-Unit $199.8M / Residential & Other $92.0M. Revenue is recognized over time as projects progress; the leading indicator is Backlog.

Contract structure / key terms. "Substantially all" contracts contain termination-for-convenience clauses — the customer can cancel, but must pay for work performed plus contractual cancellation costs. Sterling also runs construction joint ventures (e.g. the >$500M semiconductor-fab first phase is a JV ). No take-or-pay; this is backlog-and-burn, not recurring revenue.

The transformation story (matters for every other lens). Pre-2015 Sterling was a low-bid heavy-highway paver with ~4% gross margins. Since 2016 management deliberately exited low-margin work, acquired into higher-value segments, and rebranded to "Sterling Infrastructure" in 2022. The result: FY2025 consolidated gross margin 23.0% (vs 20.1% in 2024) — extraordinary for a contractor.

Lens 2 · Supply Chain

Sterling sits in the midstream of the data-center / chip-fab value chain — it is the firm that turns dirt and drawings into a power-ready, pad-ready site, and (post-CEC) wires the building.

Upstream inputs → Sterling → end customer:

• Raw materials / inputs: cement, aggregate, concrete, liquid asphalt, steel/rebar, fuel, and (post-CEC) electrical gear — switchgear, conduit, cabling. These are commodity inputs; on fixed-price work, input inflation is Sterling's risk (a named risk factor).
• Labor: project managers, estimators, superintendents, operators, electricians; partly multiemployer-union (a withdrawal-liability and bonding dependency).
• Bonding capacity (surety) is a hard upstream chokepoint specifically for Transportation — inability to bond caps the dollar value of DOT work Sterling can pursue.
• Sterling itself = site development (clearing, earthwork, utilities, foundations) + electrical/mechanical (CEC).
• Direct customers: the data-center developers and their GCs — hyperscalers, colocation providers, and "blue-chip end users" in data center, manufacturing, e-commerce, warehousing, power generation. Also semiconductor fabs and state DOTs (Transportation) and homebuilders (Building).
• End demand driver: AI/cloud compute capex. Hyperscaler 2026 capex is running ~$600B+ (Amazon, Microsoft, Google, Meta each >$100B) — Sterling is a leveraged sub-supplier to that wave.

Chokepoints / single-source dependencies: (1) Customer concentration in E-Infra — top-4 customers = 27% of segment revenue; (2) bonding for Transportation; (3) skilled-labor availability in boom regions; (4) geographic concentration (Southern/Mid-Atlantic/Rocky Mountain US + Pacific Islands). Names-or-it-didn't-happen caveat: Sterling does not disclose individual hyperscaler customer names or per-customer % — n/a — not disclosed at the named-customer level; the chain is "hyperscalers/colos" generically.

Lens 3 · Competitive Advantages (moats)

The moat is narrow but real, and entirely execution-based — there is no IP, no network effect, no switching-cost lock-in. What Sterling has:

1. Schedule assurance / on-time delivery as differentiation. For a hyperscaler, a data center one quarter late is a catastrophe (the GPUs are already bought). Sterling's stated edge is "the ability to consistently provide on-time delivery and schedule assurance through successful project management and execution… a key source of competitive differentiation, and a factor in our margin profile". This is a reputational/relational moat — you win the next phase because you didn't blow the last one. The >$1.3B "high-probability future-phase" pipeline is evidence of it.
2. Mid-market positioning. Sterling explicitly targets work "too large for small local contractors yet too small for the large national/international" players. It deliberately competes where the giants (Quanta, EMCOR) don't bother and the locals can't scale.
3. Margin discipline ("the Sterling Way"). A cultural moat: disciplined bidding, walking away from low-margin work. Backlog gross margin 17.8% at YE2025 (up from 16.7%) — it is choosing profitable backlog.
4. Scale-in-niche + alternative-delivery expertise — design-build and progressive-design-build relationships with repeat customers.

Bargaining power. Mixed. Over customers: improving — schedule-critical hyperscalers need proven executors and Sterling's margins prove pricing power. Over suppliers: weak on commodities (price-taker on cement/steel/copper). The termination-for-convenience clause means the customer ultimately holds the cancel button, capped by a make-whole. Verdict: a quality moat for a contractor, but a contractor's moat — it must be re-won every project, and it does not survive a demand collapse.

Lens 4 · Segments

All figures (FY) and (Q1):

Q1-2026 segment mix: E-Infra $597.7M (72%), +174% YoY, op margin 22.4%; Transportation $132.9M (16%), op margin 11.1%; Building $95.1M (12%), op margin collapsed to 6.5% (from 13.4%).

The story the segments tell:

• E-Infrastructure is the whole thesis. It is now ~60% of revenue (72% in Q1), grows fastest, and earns the best margin. The Q1 +174% jump includes a full quarter of the CEC electrical/mechanical acquisition. Top-4 customer concentration is falling (40%→31%→27%) — Sterling is diversifying its data-center customer base even as the segment explodes. That is a genuinely bullish structural detail.
• Transportation quietly fixed itself. Op margin 6.5% → 12.1% YoY — disciplined-bidding + IIJA/state DOT funding. The legacy "4% paver" is gone.
• Building is the soft spot. Residential concrete/plumbing hit by homebuyer affordability; commercial volume down. Margin halved in Q1. It is now the smallest, lowest-quality segment — a candidate for divestiture under "the Sterling Way."
• Geography: ~all US; E-Infra in Southeast/Northeast/Mid-Atlantic/Rocky Mountain; Building concentrated in Texas/Phoenix; Transportation in UT/AZ/CO/NV/TX + Pacific Islands.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print = Q1-2026, filed 2026-05-05)

A blowout that beat hard and triggered a guidance raise. All `` unless noted.

• Revenue $825.7M, +91.6% YoY (from $430.9M). Vs consensus ~$609.7M → +35% revenue surprise.
• Gross profit $194.3M; gross margin 23.5% (vs 22.0% PY-Q1). Margin expanding through a doubling of revenue — the CEC mix is accretive, not dilutive.
• Operating income $137.8M, +146% (from $56.1M); op margin 16.7% (vs 13.0%).
• GAAP diluted EPS $3.09 (vs $1.28 PY-Q1) — 2.4x. Adjusted diluted EPS $3.59, +120%, vs consensus ~$2.28 → +57% EPS surprise.
• Segment driver: E-Infrastructure +174% (CEC + organic data-center). Building dragged (margin 6.5%).
• Guidance — RAISED hard. FY2026 revenue to $3.7–3.8B (>50% growth over 2025's $2.49B) and adjusted diluted EPS $18.40–19.05 — midpoint +36% vs prior guide and +72% over 2025. The earlier Feb guide had been ~$3.05–3.20B / "~25% growth" — so Q1 forced a step-change.
• Backlog. Signed backlog $3.80B at 3/31/26 (vs $3.01B at YE25), book-to-burn 2.1x for the quarter; combined backlog $5.2B, +131% with E-Infra >90% of signed backlog.
• Balance-sheet flags: Goodwill $584.8M + other intangibles $548.0M = ~$1.13B intangibles post-CEC, against ~$1.1B equity. Net contract-capital builds as projects scale (working-capital headwind). Cash flow strong (FY25 op cash $440.0M ).
• Market reaction: the stock has run to all-time highs (~$1,005 52-wk high 6/4/26) and trades ~$894 — the market priced the beat and then some.
• Unusual vs own history: a +92% revenue quarter for a 35-year-old contractor is not normal; it is the inflection of a deliberate decade-long repositioning colliding with the AI capex super-cycle.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the research-layer shelf (transcripts/ empty) — sourced from web; label ``.

Tone has gone from confident (2024) to emphatic (2026):

• Recurring phrases: "data center," "mission-critical," "disciplined bidding," "schedule assurance," "the Sterling Way," "high-probability future-phase pipeline," "book-to-burn." Management consistently reframes the company as E-Infrastructure-first, deliberately downplaying the legacy paver identity.
• Trajectory of the data-center narrative: Q3-2024 "data center +~90%, >50% of segment backlog" → Q1-2025 ">65% of E-Infra backlog" → Q1-2026 ">90% of signed backlog". The concentration into one theme is rising in management's own framing — bullish for momentum, a red flag for diversification.
• New language in 2026: explicit semiconductor-fab wins (>$500M first phase, JV) and a $6.5B total addressable pool of work (signed + unsigned + future-phase). They have started quantifying the runway, which a confident management does.
• What they stopped saying: less about Building Solutions / residential — quietly de-emphasized as it weakens.

Net: management sentiment is as bullish as it has ever been, backed by numbers, but increasingly mono-thematic (it is now an AI-data-center story with a transportation tail).

Lens 7 · Comps

Peer set = the E&I / specialty-infrastructure contractors in Sterling's own proxy peer group plus the closest data-center-construction names. Multiples are `` with date, or n/a. Market caps and multiples as of mid–late June 2026.

Dividend yield / 5-yr avg ROE: STRL pays no dividend ("never paid… do not anticipate," Credit Agreement restricts). ROE is very high and rising — FY2025 net income to common $290.2M on $1,106.4M equity = ~26% ROE, and on average equity higher still; the 5-yr trend is sharply up as margins re-rated. Precise peer 5-yr ROE n/a.

Read: STRL trades at a clear premium to the E&I group (IESC/EME at ~18x EV/EBITDA, ~23–26x fwd P/E) and in line with the most expensive peer, PWR (~40x EV/EBITDA, ~50x fwd) — despite being a fraction of PWR's size and far more single-theme. The premium is defensible on margin and growth (STRL's 23% gross margin and 72% EPS-growth guide top the group) but leaves zero margin of safety if growth normalizes toward the peers'.

Lens 8 · Stock-Price Catalysts (last ~5 years)

Mostly ``. The five-year tape is one of the great industrial re-ratings: $100 invested 12/31/2020 → $1,645 by 12/31/2025 vs $188 for the DJ US Total Return Index. Total shareholder return ~989% over three years; +112.8% over the trailing 12 weeks into June 2026. 52-week range $191.00 (6/10/25) → $1,005.68 (6/4/26).

What actually moves STRL >5%:

1. Earnings + guidance raises — the dominant catalyst. Q4-2025 and Q1-2026 prints each drove sharp moves; guidance raises (the 36% EPS-guide step-up in May 2026) are the single biggest up-catalysts.
2. Data-center / AI-capex sentiment — STRL trades as a high-beta proxy for hyperscaler capex; AI-infra risk-on/off swings it. The data-center backlog disclosures are themselves catalysts.
3. Backlog & big-contract announcements — the >$500M semiconductor-fab JV, book-to-burn prints.
4. M&A — the CEC ($562M) deal repriced the growth algorithm.
5. Valuation/"overvalued" calls — pullbacks of −4% on GF-Value "still overvalued" notes show the stock is sensitive to multiple-compression chatter at these levels.

Pattern: the market reacts to forward growth signals (guidance, backlog, AI-capex), not to the balance sheet. This is a momentum/expectations stock — which cuts both ways.

---

Phase C — Judge people & books

Lens 9 · Management

• CEO Joseph A. (Joe) Cutillo — joined Sterling 2015, CEO since Feb 2017 (~8.5-yr tenure). The architect of the entire transformation.
• Track record (quantified): took a ~4%-gross-margin highway paver and built a 23%-gross-margin mission-critical infrastructure firm; oversaw EPS from $4.44 (2023) → $8.27 (2024) → $9.38 (2025) diluted and TSR of ~+1,500% over five years. By any measure one of the best small-cap industrial CEOs of the decade.
• Skin in the game: owns 1.47% ($200M) of the company. Meaningful and aligned, though not founder-level control.
• Capital allocation: disciplined and value-creating — exited low-margin work; bolt-on M&A into higher-value niches (CCS plumbing, RHB structure, CEC electrical/mechanical); buybacks ($74.2M repurchased in 2025; new $400M authorization Nov 2025 replacing the prior $200M program); no dividend (reinvest + buy back). ROE/ROIC trend strongly up on his watch.
• Archetype: professional turnaround operator, not a founder. For this stage (scaling into a capex super-cycle) that is arguably ideal — operational rigor over founder vision.
• Red flags (watch, not disqualifying): (1) Related-party leases — several principal properties (Irving TX, Flanders NJ, Draper UT, Wylie TX) are leased from related parties (see Note 19). Common in roll-ups of family-owned contractors, but a governance item to monitor. (2) Insider selling — insiders sold ~$47.6M in the trailing three months with no buying; rational at all-time highs / ~80x earnings, but not a vote of "it's cheap." (3) Earn-outs on acquisitions (CEC up to $80M) create incentive complexity.

Lens 10 · Forensic Red Flags

Acting as a forensic equity analyst. All figures `` from filings unless noted.

Income statement.

• Percentage-of-completion / over-time revenue recognition is the structural risk for any contractor — revenue depends on management's cost-to-complete estimates; aggressive estimates pull profit forward. Sterling discloses the standard "costs can vary substantially from original estimates" language and carries unapproved change orders / claims as variable consideration. No restatement history; SEC found zero LR/AAER (below). But this is the line to watch every quarter.
• Adjusted vs GAAP gap: Q1-2026 GAAP diluted EPS $3.09 vs adjusted $3.59. The add-backs (intangible amortization, acquisition costs, earn-out expense) are CEC-driven and defensible, but the ~16% adj-over-GAAP wedge will persist while CEC amortizes — non-GAAP flatters the headline.

Balance sheet.

• Goodwill $584.8M + intangibles $548.0M ≈ $1.13B, ~equal to total equity ($1.1B). CEC ($562M, much of it intangible) is the bulk. A future impairment is the single largest non-operational risk to book value if data-center demand rolls over — the 10-K explicitly flags goodwill/intangible write-down risk.
• Contract assets/liabilities: contract liabilities (billings ahead of work) $652.4M at YE25 — a source of cash (customer-financed), healthy. Watch for reversal if backlog growth stalls.
• Receivables outrunning revenue: FY25 AR swing −$171.0M (use of cash) as projects scaled — explained by larger/longer E-Infra projects, not a red flag yet, but the trajectory bears watching as projects get bigger.
• VIEs / JVs: consolidated VIEs (AP $22.1M, contract liabilities $89.6M related to VIEs) and the 50%-owned RHB now on equity method add some opacity to the consolidated picture.

Cash-flow vs earnings.

• FY25 net income $309.7M vs operating cash flow $440.0M — OCF > NI, which is reassuring (earnings backed by cash). Investing outflow $551.9M (incl. CEC cash $443M) drove the $273.5M cash decline — acquisition, not operational bleed.
• One-time items to normalize: 2024 included a $91.3M gain on deconsolidation of RHB — boosting 2024 reported pre-tax income; ensure YoY EPS comparisons exclude it (the 2025 +growth is cleaner than the raw GAAP line suggests since 2024 had this one-timer).

Stock-based comp: modest; anti-dilutive shares excluded from diluted EPS; share count basically flat (~30.9M diluted) — no SBC dilution problem.

Regulatory findings (required).:

• SEC Litigation Releases: none found naming Sterling Infrastructure since 2021 (EDGAR EFTS, LR).
• AAERs: none found (EDGAR EFTS, AAER).
• 10-K Item 3 (Legal Proceedings): Sterling discloses only ordinary-course litigation incidental to construction (incl. JVs and the 50%-owned sub); "in the opinion of management… not material". No material litigation disclosed.
• Non-SEC enforcement (web): no material FTC/DOJ/FDA/EPA enforcement actions, consent decrees, or penalties surfaced for "Sterling Infrastructure". (Note: routine environmental/RCRA exposure as a site-development contractor, but no enforcement of note.)
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-06-24. The forensic risk here is accounting-judgment (POC + goodwill), not fraud or enforcement.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (FY2026 → FY2028)

Built bottom-up from FY2025 actuals + management's raised FY2026 guide. Output `` with arithmetic shown. No forecast.ts forecast logged — --watchlist rule (breadth mode produces dossiers only).

Anchor (actuals/guide):

• FY2025 GAAP diluted EPS $9.38; adjusted ~$10.x.
• FY2026 management guide: revenue $3.7–3.8B, adj diluted EPS $18.40–19.05. Signed backlog $3.80B + combined $5.2B + book-to-burn 2.1x underwrite it.

Base case:

• FY2026 adj EPS ≈ $18.75 (guide midpoint). GAAP lower (~$16–17) given CEC amortization wedge.
• FY2027 adj EPS ≈ $22–24.
• FY2028 adj EPS ≈ $25–28.

Bull case: AI-capex super-cycle persists + share gains + more CEC-style M&A → FY2027 adj EPS ~$26, FY2028 ~$32.

Bear case: data-center capex digestion in 2027 + Building drag + margin mean-reversion → FY2027 adj EPS ~$17–19 (flat-to-down vs 2026). This is the scenario the multiple is not priced for.

Valuation sanity at ~$894: at FY2026 adj EPS $18.75 → ~48x current-year; at base FY2027 $22.5 → ~40x; at bull FY2028 $32 → ~28x. Even on the bull two-years-out number the stock is ~28x — i.e. the price already capitalizes several years of flawless execution. GF Value intrinsic estimate is ~$207 (vs ~$894 price → stock ~331% above it) — an extreme but directionally honest flag.

Lens 12 · Bull vs Bear

Bull case. Sterling is the highest-quality, highest-margin pure-play on the single biggest industrial capex wave of the era — the physical buildout of AI compute. It has (1) a proven, disciplined operator compounding EPS ~5x in three years; (2) a 2.1x book-to-burn and a $6.5B addressable work pool giving multi-year visibility; (3) margin expansion through scale (gross margin up while revenue doubles) — the opposite of the usual contractor decremental; (4) the CEC electrical/mechanical addition moving Sterling up the value chain into higher-margin, higher-skill work right as power is the data-center bottleneck; (5) a clean balance sheet (net cash-ish ex-CEC, OCF > NI, no dividend drag) and a $400M buyback. If hyperscaler capex stays at $600B+ and Sterling holds share, the FY2026 guide ($18.40–19.05 adj EPS) proves conservative and FY2027–28 compounds — a 50x multiple on 70% growth is not absurd while the growth lasts.

Bear case (permanent-impairment risks).

1. Single-theme, cyclical, customer-concentrated demand. E-Infra is >90% of signed backlog and data-center is the bulk of it; top-4 customers = 27% of segment. A hyperscaler capex digestion (history says these waves over-build, then pause) hits Sterling's growth engine directly and at once. Termination-for-convenience means backlog can soften faster than it built.
2. Valuation = no margin of safety. ~80x trailing / ~50x forward / ~331% above GF intrinsic. The stock has already re-rated from ~17x (5-yr median) to ~50x. Multiple compression alone — back to even a still-rich 30x on flat-to-up EPS — is a ~40% drawdown.
3. Goodwill/intangible (~$1.13B ≈ equity). A demand rollover forces a CEC impairment, gutting book value and confirming the cycle turned.

Pre-mortem (18 months out, thesis broke): It's late 2027. Hyperscalers guided 2027 capex flat-to-down after a 2024–26 over-build; Sterling's book-to-burn fell below 1.0x; two large data-center phases were "paused for convenience"; Building stayed weak; the Street re-rated STRL from 50x to 25x on decelerating growth and the stock halved from its highs even as EPS held — a textbook great-business-bad-entry-price outcome.

Is the multiple too high? On any normalized/through-cycle basis, yes — it prices perfection. On a momentum/next-12-months basis, the guidance raise and 2.1x book-to-burn justify a premium, just not necessarily this premium.

Contrarian view (what the market refuses to see): Bulls treat the data-center backlog concentration as pure strength; it is also the fragility — the same single theme that drove the +1,500% TSR is the single point of failure, and the market is extrapolating a capex flow (annual buildout) as if it were a recurring stock. The unglamorous Transportation segment (margin doubled, IIJA-funded, far more durable) is the part of Sterling that survives a data-center pause — and it's the part nobody is paying for.

Lens 13 · Devil's Advocate (short-seller)

Skeptical short-seller dismantling the bull case:

• What structurally breaks the model: Sterling makes money by winning the next phase of mega-projects from a handful of hyperscalers. That is a flow business dressed as a franchise. When the AI buildout shifts from "land-grab" to "digest," book-to-burn cracks below 1.0x and the revenue that doubled can halve — there is no recurring revenue, no contract that the customer can't cancel for convenience.
• Revenue concentration: E-Infra >90% of signed backlog; top-4 customers 27% of that segment. Lose one mega-hyperscaler's pipeline and the growth algorithm inverts overnight. Geographic concentration (a few US regions) compounds it.
• Why the moat is weaker than bulls think: there's no IP, no switching cost, no network effect — only reputation and schedule execution, which Quanta (PWR, $110B), EMCOR (EME), IES (IESC), Comfort Systems (FIX) and Primoris (PRIM) all also have, with deeper balance sheets. As hyperscaler work becomes the prize, the giants will "converge into the mid-market" — Sterling's own 10-K warns of exactly this margin-compressing convergence.
• Most dangerous competitor bulls underestimate: Quanta + EMCOR on the electrical/mechanical side — post-CEC, Sterling is now competing for the high-margin power scope against firms 4–10x its size with national labor pools. CEC may have bought Sterling into its most competitive niche at peak.
• Worst capital-allocation / governance items: related-party leases on key properties (Note 19); $47.6M insider selling at the highs; an $80M CEC earn-out that can distort near-term operating-income incentives; ~$1.13B of goodwill/intangibles to defend.
• Assumptions that must hold for $894: ~70% EPS growth this year and a 50x multiple and multi-year data-center capex at $600B+ and CEC integrating cleanly and margins holding while revenue compounds. Miss any one and the re-rating reverses.
• If growth disappoints 20–30%: FY2027 adj EPS ~$17–19 instead of ~$22–24, and the multiple compresses from ~50x toward the peer ~25–30x → a 40–55% drawdown is entirely plausible without anything "going wrong" operationally.
• Single scenario that permanently impairs: a sustained AI-capex retrenchment (compute glut / hyperscaler ROI scare) that turns the data-center buildout from secular-growth to over-built-pause — forcing a CEC goodwill impairment and re-rating Sterling back to a cyclical contractor multiple. Plausibility: not base-case in 2026, but a real tail by 2027–28, and the price gives you no cushion for it.

Lens 14 · Management Questions (ordered by information value)

1. What is your book-to-burn assumption embedded in the FY2026 guide, and at what book-to-burn level does revenue growth go flat — i.e., how much cushion is in the $5.2B combined backlog if no new mega-awards land in 2H26?
2. How concentrated is E-Infrastructure backlog by end customer — what % is the top customer and top three, and how has that changed as the segment scaled to >90% of signed backlog?
3. What is your direct read on hyperscaler 2027 capex intentions from your customers' multi-year deployment plans — are you seeing any phase pauses, push-outs, or "digestion" language yet?
4. On CEC: what was the trailing revenue/EBITDA you paid $562M for, what's the integration status, and how should we think about electrical/mechanical margins versus your legacy site-development margins through a cycle?
5. What share of the >$500M semiconductor-fab JV revenue/margin accrues to Sterling, and how does the JV structure change the risk profile versus a wholly-owned contract?
6. Where can consolidated gross margin realistically go from 23%, and what mix shift (E-Infra share, electrical vs site-dev, CEC) drives it — or is 23% the peak?
7. What is your capital-allocation priority stack now — organic capex, M&A, or the $400M buyback — and at ~50x forward earnings, how do you weigh repurchasing your own stock?
8. Building Solutions margin fell to ~6.5% — at what point does "the Sterling Way" logic say divest it, and what would you do with the proceeds?
9. How exposed is the fixed-price book to commodity (copper/steel/cement) and skilled-labor inflation, and what protection (escalators, pre-buy) is in place on the largest E-Infra contracts?
10. On the related-party leases (Note 19): walk us through the terms, how they're priced at arm's length, and your plan to reduce related-party dependence over time.
11. How do you defend the high-margin power/electrical scope against Quanta and EMCOR as they move into the data-center mid-market your own 10-K flags as a convergence risk?
12. What's the realistic skilled-labor ceiling on how fast you can grow E-Infra — at what revenue run-rate does labor availability, not demand, become the binding constraint?
13. How should we think about contract capital / working-capital intensity as projects get larger and longer — is the AR build in FY25 the new normal, and what does it do to free-cash conversion?
14. What is your M&A pipeline and discipline at current valuations — are private E&I assets still acquirable at multiples accretive to a 50x-earnings acquirer?
15. What single development over the next 24 months would most change your own view of the business — what keeps you up at night?

---