Cloud Computing
PrivateA top-quartile-margin, asset-light infrastructure-design compounder re-rated to a ~30% discount to peers on a cash-flow timing scare and a misplaced "AI eats engineering" fear — the disconnect (FY26 EPS guide RAISED to $5.90–$6.10 while the stock halved from $134) is the opportunity, gated on H2 free-cash-flow normalising.
Research
The verdict
A top-quartile-margin, asset-light infrastructure-design compounder re-rated to a ~30% discount to peers on a cash-flow timing scare and a misplaced "AI eats engineering" fear — the disconnect (FY26 EPS guide RAISED to $5.90–$6.10 while the stock halved from $134) is the opportunity, gated on H2 free-cash-flow normalising.
Primary sources
Source documents — open to read in full
AECOM (Dallas, TX; NYSE: ACM; FY ends the Friday nearest Sept 30) provides advisory, planning, consulting, architectural & engineering design, program management and construction management across transportation, water, environment, facilities and energy. Per ENR's 2025 Design Survey it is "the largest general architectural and engineering design firm in the world… and #1 ranked water, transportation design, facilities design, environmental engineering, environmental consulting and environmental science firm".
Business model. It is a labour-arbitrage / brain-for-hire model: revenue ≈ (billable employees × utilisation × bill rate), cost ≈ (compensation + subcontractors + other direct costs). The single most important structural fact about the income statement: ~53% of reported "revenue" is pass-through — subcontractor and other direct costs routed through GAAP revenue ($8.6B of $16.1B in FY25, down from 56% in FY24). So gross revenue overstates the real business by ~2x; the firm and the Street manage to Net Service Revenue (NSR) = revenue excluding pass-throughs (~$7.5B FY25 run-rate; $1.948B in Q2 FY26, +2% cc ). Read NSR, not revenue.
Three segments:
Customers. Heavily public-sector / government: US DOTs, Department of Homeland Security/FEMA (emergency response), Department of Defense, Department of Energy, plus private clients in industrial, data-center, aviation and commercial facilities. customers.csv is empty in the research layer — no single named customer concentration is disclosed, consistent with a fragmented, project-by-project, thousands-of-contracts book.
Contract structure (Q2 FY26 disaggregation):
AECOM's "supply chain" is a labour-and-subcontractor chain, not a physical one. Map:
Upstream inputs → AECOM → end customer
Chokepoints / single-source risks: (1) Talent — the only genuine scarce input; an engineer shortage caps growth. (2) Government funding cadence — IIJA, state DOT budgets, UK/Middle East programs; a budget freeze hits the order book (mitigated: "only 36% of IIJA funding in AECOM's markets has been spent" ). No physical/commodity chokepoint exists — this is the asset-light advantage. Lens fails generic; the named chain above is specific.
Where the moat is real:
Bargaining power. Over suppliers (subcontractors/labour): moderate — subs are many-sourced, but engineering talent has pricing power over AECOM (wage inflation). Over customers: moderate-to-weak on price (public procurement is competitively bid and fragmented — the 10-K calls the markets "highly fragmented") but moderate-to-strong on retention once embedded.
Verdict on moat: a structural-scale + switching-cost + reputation moat at the top of the market, not a pricing monopoly. Durable but not impregnable — the field is an oligopoly of ~5 credible giants, all chasing the same mega-programs. The AI question (does AI widen the moat for the scale leader, or commoditise the engineering hour for everyone?) is the central debate and is unresolved.
segments.csvis empty in the research layer, so all segment numbers below are pulled directly from the filings and labelled.
FY2025 (full year), by reportable segment — gross-revenue basis:
| Segment | Revenue | YoY | Gross profit | GP margin | YoY GP margin |
|---|---|---|---|---|---|
| Americas | $12,525.9M | +0.3% | $882.1M | 7.0% | +0.9pp (from 6.1%) |
| International | $3,613.2M | −0.1% | $334.1M | 9.2% | +0.3pp (from 8.9%) |
| AECOM Capital | ~$0 | — | (JV-driven) | — | — |
| Total | $16,139.6M | +0.2% | $1,216.7M | 7.5% | +0.8pp (from 6.7%) |
Two structural reads: (a) International runs a higher gross margin (9.2%) than Americas (7.0%) because Americas carries proportionally more low-margin pass-through construction-management revenue — i.e. Americas' gross margin understates its net (NSR) margin. (b) The whole-company gross margin expanded ~80bps YoY on flat revenue — proof the engine is margin-expansion, not volume, driven by restructuring benefits, offshore Enterprise Capability Centers, continuous-improvement and mix-shift to higher-margin advisory.
End-market dynamics within Americas FY25: Transportation +$261.1M (+11.7%), Water & Environment +$127.4M (+6.0%), Facilities −$344.7M (−4.3%, the soft spot, partly pass-through roll-off). Transportation and Water are accelerating; Facilities decelerated in FY25 but is the line that captures the data-center / high-tech demand now inflecting.
Q2 FY26 segment "earnings before income taxes & amortization" (new ASU-2023-07 disclosure): Americas $233.9M (+9.9% YoY), International $82.5M (+9.4%), ACAP −$1.5M. On a NSR + adjusted-operating-margin basis (how the Street tracks it) Q2 FY26: Americas NSR $1.2B at a record 20.0% adj. operating margin; International NSR $754M at 11.1%. Americas at a 20% margin on NSR is genuinely best-in-class for the sector.
Geographic disaggregation Q2 FY26: Americas $2,911.6M, EMEA-India-Africa $545.8M, Asia-Australia-Pacific $343.8M.
The print itself:
Beat/miss & reaction (the important part). Adjusted EPS beat, margins hit records, and guidance was raised for a second straight quarter (see below) — yet the stock FELL on the print. The market fixated on two GAAP-level negatives: (1) GAAP revenue ~flat and below the ~$4.01B "expected" figure; (2) free cash flow −$27.4M vs +$178.4M a year earlier. The 10-Q corroborates the cash drag: H1 FY26 operating cash flow collapsed to $74.0M from $341.7M, with cash down $551.4M to $1,034.3M, driven by a −$282.1M working-capital build and "a longer than anticipated claims resolution process".
Guidance & tone. Management RAISED FY26: adjusted EPS to $5.90–$6.10 (from $5.85–$6.05; +14% YoY at midpoint), adjusted EBITDA $1.275–$1.305B (+7%), segment adjusted operating margin ~16.8%, organic NSR growth 6–8%, and reiterated ~$400M free cash flow for the year. Tone was confident — "record backlog… double-digit pipeline growth."
Balance-sheet flags. Net leverage 1.2x. Interest expense rising ($50.6M Q2 vs $42.2M; $95.8M H1 vs $85.2M) as the 5%+ 2033 notes replaced cheaper debt. H1 discontinued-ops loss of $70.2M — legacy self-perform construction wind-down, not continuing operations (see Lens 10), but it dragged GAAP H1 net income.
Unusual vs. its own history: the negative quarterly FCF is the genuine anomaly. AECOM's cash flow is structurally H2-weighted, and management reaffirmed the ~$400M full-year FCF, so this is most likely a timing/claims issue rather than structural — but it is the single most important thing to verify at Q3 (early Aug 2026).
transcripts/is empty in the research layer; call commentary below is ``.
The narrative arc across recent calls is remarkably consistent and increasingly confident:
Peer set = the global infrastructure-design oligopoly. Multiples are `` June 2026 snapshots; where a clean figure was not sourced it is marked n/a (no fabrication).
| Company | Ticker | Mkt cap | Trailing P/E | EV/adj EBITDA (FY26e) | Div yield | Notes |
|---|---|---|---|---|---|---|
| AECOM | ACM | ~$9.1B | ~14.7x | ~9.5x | ~1.5% (≈$1.04 ann.) | Top-quartile margin, cheapest multiple |
| WSP Global | WSP.TO | ~C$29.3B | ~24x | n/a | low | Premium-rated, asset-light design pure-play |
| Jacobs Solutions | J | ~$14.4B | ~38x (optically high post-CMS spin) | n/a | low | Spun off govt-services (Amentum); now design-led |
| Stantec | STN | ~$7.9B | ~22x | n/a | low | Pure design, "pristine balance sheet" |
| Tetra Tech | TTEK | n/a (mkt cap) | n/a | n/a | ~1.0% | Water/environment-weighted; EBITDA margin +90bps Q2 FY26 |
| AtkinsRéalis | ATRL.TO | n/a | n/a | n/a | low | Structurally weaker margin consistency vs peers |
Peer-group valuation summary (the load-bearing comp): AECOM trades at ~9.5x EV/adj EBITDA on FY26 estimates vs ~13.3x for the peer set — a ~30% discount — despite a 16.8% TTM segment adjusted operating margin that exceeds the ~15.2% peer average (peer set = Jacobs, Tetra Tech, Stantec, WSP). On P/E the gap is starker still: ~14.7x vs WSP ~24x / Stantec ~22x / Jacobs ~38x. AECOM is the highest-margin and the cheapest name in its own peer group simultaneously — the definition of a valuation anomaly, if the cash flow and AI fears are transitory.
5-year average ROE: n/a cleanly across all peers (would require a consistent multi-year pull). AECOM's own returns trend is strongly positive — management cites the buyback-driven 19% share-count reduction and rising margins; ROE is structurally high for an asset-light model.
[All ``.] AECOM's price reacts to margin/guidance and capital-return signals far more than to revenue, and recently to sector-wide AI sentiment:
Pattern read: this name trades on (a) margin/EPS guidance, (b) capital-return cadence, and (c) sector AI narrative — not on gross revenue (which is half pass-through and which management is deliberately shrinking). The current dislocation is a sentiment/cash-timing de-rate layered on top of rising fundamentals.
CEO — W. Troy Rudd (CEO since Aug 2020; previously CFO from Sept 2015). A finance-first operator installed during the Starboard Value activist era. His track record is the transformation itself: exited self-perform at-risk construction (civil/power/oil&gas), shed the Management Services segment, slashed the risk profile, drove gross margin from ~6% to 7.5%+ and segment adj. operating margin to a record ~16.5% while returning $3.5B+ to shareholders since Sept 2020 and cutting the share count ~19%. That is a credibly value-creating tenure — the company today is a cleaner, higher-margin, asset-lighter business than the one he inherited.
President — Lara Poloni (promoted from CE of EMEA) — operational counterweight to Rudd's finance background.
Tenure & skin in the game. Insiders (officers + board) hold <2% of shares; institutions (BlackRock, Vanguard et al.) dominate. Low insider ownership is typical for a post-activist large-cap professional-services firm but is a mild alignment negative — management is an agent, not a large owner. insider-transactions.csv is absent from the research layer.
Capital allocation — the strongest part of the story. Explicit "returns-driven capital allocation policy… return substantially all available cash flow to stockholders". Track record: $3.5B+ returned since 2020 (~$1.8B+ buybacks alone, −19% shares); $1.0B buyback authorization (Nov 2024), $644.4M remaining at FY25-end; dividend raised 18% to $0.26/qtr ($1.04 ann.), having started at $0.15 in 2022. Bolt-on M&A is small and disciplined (2 deals FY25, $212.5M; "not material to consolidated results") — no empire-building. This is a management team that treats the share count as something to shrink.
Founder vs professional manager. Pure professional managers (no founder); appropriate for a mature, scaled, capital-return-stage business. The risk is the flip-side: professional managers optimise the financials they're measured on — which is exactly why the FCF wobble matters (Lens 10).
Red flags: none material on comp or related parties surfaced. The historical governance event is the 2020 Starboard board fight: Starboard (Peter Feld, ~3.7–4% stake) opposed Rudd's selection; three directors voted against; Feld resigned from the board in protest. Chair/CEO roles were split (Douglas Stotlar chairs) — a governance positive. Net: a board reformed by activism, now stable.
Acting as a forensic analyst. AECOM is a percentage-of-completion / over-time revenue recogniser — the classic accounting-risk profile for E&C — so the scrutiny goes to estimates, claims and cash conversion.
Regulatory findings (required sub-section):
Built bottom-up from management's reaffirmed FY26 guidance + the structural margin-expansion algorithm. Output ``; inputs labelled.
Anchor (FY26, from raised guidance): adjusted diluted EPS $5.90–$6.10 (midpoint $6.00); adj EBITDA $1.275–$1.305B; segment adj op margin ~16.8%; organic NSR growth 6–8%; FCF ~$400M.
The EPS algorithm = organic NSR growth (mid-to-high single digit) × steady operating leverage (margin +30–50bps/yr) × buyback-driven share-count reduction (~3%/yr) − rising interest. This has reliably produced low-double-digit adjusted-EPS growth.
| FY26e (anchor) | FY27e | FY28e | |
|---|---|---|---|
| Bear | $5.90 (guide low) | $6.30 | $6.70 |
| Base | $6.00 (guide mid) | $6.75 | $7.55 |
| Bull | $6.10 (guide high) | $7.05 | $8.15 |
Cross-check vs valuation: at ~$68 the stock is ~11.3x FY26 base EPS ($6.00) and ~10.1x FY27 base ($6.75). For a company guiding 14% EPS growth with record margins and a 1.2x book-to-burn, a re-rate to even ~14–15x FY27 (peer-discounted, not peer-parity) implies ~$95–$100 — roughly the price the seed targets. The bear's "$15B revenue / −6%" framing is a gross-revenue artefact of shedding pass-through; on NSR the company is growing 6–8%.
Tracked forecast: Skipped the forecast.ts create step per --watchlist rules (log a Brier forecast only on genuine, human-gated commitment). Candidate to log at /thesis: ACM FY27 non-GAAP EPS >= $6.50, p≈0.62, resolves 2027-09-30.
Bull case. AECOM is the highest-margin, asset-light, capital-return-disciplined leader of the global infrastructure-design oligopoly, trading at a ~30% EV/EBITDA discount to lower-margin peers. The growth is structural and funded by other people's budgets: a multi-decade super-cycle of (1) transportation (record state DOT budgets; only ~36% of IIJA spent ), (2) water (drought, contaminants, ageing systems), (3) environment (permitting, remediation, energy transition), and (4) the AI-infrastructure build-out — AECOM designs the power, water, cooling and facilities that data centers need ("built over 11 GW across 45 countries over 25 years… hyperscale data centers, semiconductor fabs, liquid-chilled cooling, advanced water treatment" ). The reflexive AI moat: scale → proprietary data → better AI design tools → faster delivery and higher margin (the contrarian "AI moves the margin needle" thesis ). Capital allocation is a relentless tailwind — ~3% of shares retired annually, dividend +18%. Earnings surprise potential: an H2 FCF snap-back that kills the cash-flow bear in one print.
Bear case (what could permanently impair, or just keep the multiple low).
Are multiples too high? No — the opposite. At ~11x FY26 EPS / ~9.5x EV/EBITDA for a 16.8%-margin, mid-teens-EPS-grower, the multiple is too low versus both peers and the company's own history (it traded near 2x this multiple at the Oct-2025 high). The risk is a value trap (multiple stays low) more than overvaluation.
Contrarian view (what the market refuses to see): the Street is pricing AI as a revenue threat to AECOM when the evidence is that AI is a margin tailwind for the scale leader — and is simultaneously punishing a deliberate shrinking of low-margin pass-through revenue as if it were organic decline. The market is conflating "gross revenue down" with "business shrinking," when NSR is growing 6–8% at record margins.
Dismantling the bull case.
The default arms dealer of the AI buildout — a real moat compounding a $15B backlog into 30% organic growth, but priced at 82x for perfection while insiders sell 65:0 and EMEA orders are already cracking.
A 109-year-old radiator maker that 80/20'd itself into a data-center liquid-cooling growth story and got re-rated 14.7x in five years — the asset is real and a single hyperscaler has pre-committed >$4B of cooling for CY27–29, but at ~38x forward / ~133x trailing the price already pays for flawless execution while Q4 just showed the margin and supply-chain cracks that flawless execution doesn't have. Own the post-RMT pure-play on a data-center capex scare or a second margin miss — not here.
A quietly excellent RF-and-photonics compounder that the market has finally noticed — fab-lite 56% gross margins, three secular legs (defense, AI data-center optics, telecom), a net-cash balance sheet after killing its 2026 converts, but now priced at ~54× forward with an AI-optical multiple it must keep compounding into; great business, demanding entry.