Jacobs Solutions

Phase A — Understand the business

Company Overview

Jacobs Solutions Inc. (Dallas, TX; NYSE: J; ~75-year-old firm, ~43,000 employees, operating in 40+ countries) is a professional-services company delivering engineering, design, architecture, construction management, O&M, and digital/technical consulting. It earns revenue under cost-reimbursable and fixed-price contracts across a diversified base of government, industrial and commercial clients — it is a labor/expertise business, not a capital-asset business (capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs. was only ~$79M on $12.0B revenue in FY2025 — an asset-light model).

Two reportable segments post-reorganization:

• Infrastructure & Advanced Facilities (I&AF) — ~89% of revenue. The core E&C engine: water, transportation, environmental, energy-transition, life-sciences, semiconductor, and advanced facilities including data centers. FY2025 external revenue $10,764.2M.
• PA Consulting — ~11% of revenue, but the high-margin crown jewel. A UK-based management/strategy/innovation consultancy. FY2025 external revenue $1,265.6M. As of March 2026, 100% Jacobs-owned (see Lens 5).

The transformational event: On Sept 27, 2024, Jacobs completed a Reverse Morris Trust spin-off of its Critical Mission Solutions (CMS) + Cyber & Intelligence government-services businesses, merging them with Amentum (Jacobs holders got 51% of Amentum; Jacobs retained a ~7.5% stake). This deliberately shed lower-margin, gov-heavy work to become a "simpler, more focused, higher-margin" company. Result: US federal government fell to just 8% of FY2025 revenue (from 10% FY2024) — Jacobs is now overwhelmingly a commercial-and-infrastructure services firm, not a defense/gov contractor.

Contract structure: Mix of cost-reimbursable (lower-risk, cost-plus) and fixed-price (margin upside but execution risk). Backlog $23.1B at FY2025-end (+$1.2B YoY); remaining performance obligationsRemaining performance obligationsRevenue a company has already signed contracts for but has not yet delivered or booked. (ASC 606) $15.7B, 51% recognizable within 12 months. Record backlog $27.0B (+22% YoY) reported at Q2-FY2026 — a leading indicator pointing up.

Supply Chain

For a services firm, the "supply chain" is talent in → billable expertise → delivered project, plus the physical construction chain it orchestrates on behalf of owners. Named stakeholders along the chain:

• Upstream inputs: Engineering/technical labor (the binding constraint — see Lens 3), subcontractors, and procured equipment/materials that flow through as pass-through cost (the ~$3.3B gap between gross and net revenue). Jacobs is the orchestrator, not the buyer of record on most hardware.
• The company: Jacobs provides design, engineering, program management and, on marquee data-center jobs, EPCM (Engineering, Procurement & Construction Management).
• Downstream end customers (named):
• Data centers / AI infra: NVIDIA (co-developed a Data-Center Digital-Twin on the NVIDIA Omniverse DSX blueprint), Hut 8 (~$10B hyperscale AI data-center program, EPCM + digital twin), PsiQuantum (high-performance-computing environments).
• Water/transport/environment: government agencies, municipalities and utilities across the US, UK, Middle East, ANZ.
• Life-sciences / semiconductor: advanced-manufacturing owners (the "Advanced Facilities" franchise).
• Chokepoint / single-source dependency: The scarce input is specialized engineering headcount — mission-critical/data-center design talent in particular. Jacobs' edge is the installed base of that talent (decades of advanced-facility delivery); its risk is wage inflation and poaching. There is no physical single-source chokepoint (asset-light), which is a structural positive.

Competitive Advantages (moats)

Jacobs competes with AECOM, WSP Global, Stantec, Tetra Tech, Fluor, Arcadis, Mott MacDonald and others. Durable moats:

1. Scale + reference base in advanced facilities. As AI data centers scale to gigawatt size, owners want certainty before committing capital — a firm that has "delivered the world's most advanced facilities for decades" is a shortlist of a handful globally. This is a reputation/track-record moat that is hard to replicate and is exactly what is being monetized now (CEO: the DC "investment cycle is still in early stages").
2. Switching costs mid-program. Once Jacobs is embedded on a multi-year design/EPCM program (backlog $27B), swapping the engineer of record is costly and risky for the owner — sticky, multi-year revenue.
3. PA Consulting = a differentiated, higher-value moat. A management-consulting arm most pure-play E&C peers lack, letting Jacobs sell strategy + digital + engineering together at consulting-grade (~22%) margins.
4. Digital/IP layer. Proprietary tools — Data-Center Digital Twin (NVIDIA Omniverse), Evolve, Climate Risk Manager, Intelligent O&M — push Jacobs up the value chain from "hours" toward "technology-enabled outcomes".

Bargaining power: Modest. This is a competitive-procurement business; contracts are won on price, schedule and technical sophistication, and the 10-KForm 10-KA company’s audited annual report to the US regulator. The most complete thing it publishes. explicitly flags that competition "can place downward pressure on our contract prices and profit margins". Jacobs has power over subcontractors and pricing power in scarce-skill niches (mission-critical/DC), but limited power over large sophisticated buyers (hyperscalers, governments). The moat is relative (better shortlist position than most peers), not monopolistic.

Segments

All segment figures ``.

By segment — 3-year revenue & segment operating profit (external customers, $M):

Trend & cause:

• I&AF growing mid-single-digits on GAAP revenue with expanding segment margin (7.6% → 7.7% → 8.4%). FY2025 growth "mainly driven by the Company's I&AF business… stronger performance in its Advanced Facilities and APME operations". Advanced Facilities (semis, life sciences, data centers) is the accelerant.
• PA Consulting is the margin engine: ~22% segment margin vs I&AF's ~8%. FY2025 revenue +7.5% and OP +16.4% — accelerating and mix-accretive. Owning 100% of PA (from March 2026) captures more of this profit for Jacobs shareholders.
• Reconciliation to GAAP: Total segment OP $1,182.0M reconciles down to GAAP operating profit $863.6M via restructuring/transaction charges (−$162.9M) and intangible amortization (−$155.5M).

By geography (FY2025 GAAP revenue, $M): United States $7,415.5 (62%), Europe $2,871.6 (24%), Middle East & Africa $596.3, Australia/NZ $573.2, Canada $248.1, India $180.1, Asia $144.9. Heavily US/Europe weighted. US federal government = only 8% of total — the Amentum spin structurally de-risked government-shutdown / budget exposure.

---

Phase B — Measure performance

Earnings Result

The story of the last two prints is one number that looks alarming and isn't: a Q2-FY2026 GAAP operating loss that is 100% a deal-charge artifact, sitting on top of an operating business that is accelerating.

FY2025 (full year, ended 2025-09-26) — the first clean post-spin year:

• GAAP revenue $12,029.8M (+4.6% YoY); gross profit $2,984.9M (24.8% margin, +20bps); GAAP operating profit $863.6M (+24.7% YoY, 7.2% margin).
• Net earnings to Jacobs from continuing ops $313.3M, diluted EPS $2.58 — down 48.9% vs FY2024's $4.79. The decline is optical: FY2024 continuing EPS was inflated by a +$186.9M mark-to-market gain on the retained Amentum stake, whereas FY2025 booked a −$227.3M mark-to-market loss on that same stake plus a $20.5M debt-extinguishment loss. Strip the Amentum-stake noise and operating profit rose ~25%.
• Non-GAAP (the number management/Street watch): Adjusted net revenue $8.7B (+5.3%), Adjusted EBITDA $1.2B (+13.9%), Adjusted EPS $6.12 (+15.9%). Adjusted EBITDA growing ~3x faster than GAAP revenue = the margin-mix thesis working.
• Balance sheet FY2025-end: cash $1,235.4M, total debt $2,236.5M → net debt ~$1,001M. Goodwill $4,780.8M = 42% of total assets (CH2M + PA legacy) — a watch item, not yet an impairment. Operating cash flow $686.7M, down from $1,054.7M FY2024 (working-capital build in receivables, −$177.9M) — the one genuine yellow flag in FY2025.
• Backlog $23.1B (+$1.2B); $754.1M of stock repurchased with $1.22B remaining authorization.

Q2-FY2026 (quarter ended 2026-03-27) — the print that needs decoding:

• GAAP revenue $3,694.9M (+27% YoY) — the jump reflects PA now fully consolidated + organic growth. Gross profit $794.9M.
• GAAP operating LOSS $(81.2)M vs +$208.6M prior year; net loss to Jacobs $(45.9)M; diluted EPS $(0.32) continuing / $(0.34) total.
• Why: SG&A ballooned to $876.1M (+65% YoY) because Q2 absorbed ~$207.0M of incremental one-time PA Consulting Transaction charges — mainly full vesting of PA equity-incentive grants plus $123.9M of accrued PA employee-benefit-trust distributions — plus $352.2M of total restructuring/transaction/other charges below the segment line. Segment operating profit was actually healthy: I&AF $225.2M + PA $79.9M = $305.1M. The GAAP loss is entirely the deal cost; the underlying business made money.
• Proof it's Q2-specific: the six-month figures are positive — net income $79.6M, diluted EPS +$0.79.
• Non-GAAP (what the market traded on): Adjusted EPS $1.75 vs $1.63 est (+7% beat), +22.4% YoY; adjusted net revenue $2.33B (+8.8% YoY, +2% beat). Record backlog $27.0B (+22%). Data-center revenue +100%+ YoY, pipeline +400%; AI-infra ecosystem now 10–11% of the business, growing >40%.
• Guidance raised for the 2nd straight quarter: FY2026 adjusted net-revenue growth 8.0–10.5% (was 6.5–10.0%), adj EBITDA margin 14.6–14.9%, adjusted EPS $7.10–$7.35 (was $6.95–$7.30), adj FCFFree cash flowCash left after paying to run and maintain the business. Unlike profit, it is hard to flatter with accounting choices. margin 7.0–8.5%.

Market reaction: modestly positive (~+2.4% after-hours) — the adjusted beat + guidance raise + record DC backlog outweighed the headline GAAP loss, i.e. the market correctly looked through the PA charge.

Unusual-vs-history flags: (1) the two Amentum-stake marks (+$187M / −$227M) that whipsaw GAAP net income; (2) the Q2 one-time PA charges; (3) the FY2025 OCF decline on receivables build; (4) net debt roughly tripling to ~$2.7B post-PA buyout (Lens 9/10).

Earnings Calls (sentiment trend)

No transcripts on the research-layer shelf (transcripts/ empty; ingest not run per task scope). Sentiment reconstructed from filing MD&A + earnings coverage ``, so treat as directional.

• Tone: consistently and increasingly bullish, with a clear pivot to AI/data-center as the lead narrative. Management launched "Challenge Accepted" (Feb 2025 investor day) as the multi-year growth story, and by Q2-FY2026 the framing is explicitly AI-infrastructure-led.
• What they keep saying: "simpler, more focused, higher-margin company"; "positioned across the entire infrastructure ecosystem"; data-center cycle "still in early stages"; "profitable growth" + "strong balance sheet." Recurring metric emphasis on adjusted net revenue, adjusted EBITDA margin, and organic growth — a deliberate steer away from noisy GAAP.
• What they stopped saying: government/defense-mission framing (shed with Amentum) and Amentum-stake commentary (monetized/wound down — the retained equity-securities line went to $0 on the FY2025 balance sheet).
• Shift over time: FY2024 calls = "executing the separation"; FY2025 = "clean focused company, margin expansion"; FY2026 = "AI-infra supercycle, DC +100%, pipeline +400%, raising guidance." The narrative has moved from portfolio surgery to secular growth — a healthy progression, but it also means expectations (and the multiple) now embed the DC ramp.

Comps

Peer set: global engineering/consulting firms. Multiples are ``, dated; where a specific figure wasn't sourced it is marked n/a. Do not treat any cell as filing-grade.

Read: The peer group (Jacobs, Tetra Tech, Stantec, WSP) trades ~13.3x EV/adjusted EBITDA on average. Jacobs at ~14.2x sits at/slightly above the peer average and roughly in line with WSP — i.e. fully valued, not cheap. AECOM is the value outlier (9.5x FY26E) despite higher margins, which is the single most obvious relative-value flag against owning Jacobs here: a skeptic would ask why pay a premium multiple for J when ACM is cheaper on superior margins. Jacobs' premium is defensible only if its DC/AI-infra mix and PA-consulting margins compound faster than peers — the bull thesis in one line. FY26E fwd P/E on the adjusted EPS midpoint ($7.22) at $125.58 is ~17.4x, a reasonable-not-cheap multiple for a mid-teens EPS grower.

Stock-Price Catalysts (5-yr, >5% moves)

Mostly ; pattern analysis is from the reconstructed history.

• Total returns: 2021 +28.8%, 2022 −13.0%, 2023 +9.2%; ~5-yr avg ~10.5%/yr.
• The dominant multi-year catalyst = portfolio transformation, not any single earnings print:
• Dec 2017: Closed CH2M acquisition ($3.27B EVEnterprise valueWhat it would cost to buy the whole company: its market value plus its debt, minus the cash you would get with it. Often very different from market cap.) — created a ~$15B professional-services leader; the deal that built today's water/environmental scale.
• Nov 2023 → Sept 2024: Announced and completed the Amentum Reverse Morris Trust spin — the re-rating catalyst that turned J from a gov-heavy contractor into a focused, higher-margin services firm.
• Feb 2025: "Challenge Accepted" investor day (6–8% organic net-rev growth, 16%+ EBITDA margin by FY29) — reset the growth framing.
• Nov 2025 (Q4/FY25): Beat + margin expansion; but analysts cut forecasts on some line items post-print — a reminder expectations are demanding.
• Jan–Mar 2026: £1.2B PA Consulting buyout + debt financing — the event driving the Q2 GAAP loss and the balance-sheet re-leveraging.
• May 2026 (Q2): Adjusted beat, 2nd guidance raise, record $27B backlog + DC +100%/pipeline +400% — the AI-infra narrative crystallized; stock +2.4% after-hours.
• What the market actually reacts to (pattern): (1) portfolio/strategic actions (spins, big M&A) move the multiple the most; (2) backlog + organic-growth + adjusted-EBITDA-margin beats/raises move it quarter-to-quarter; (3) it looks through GAAP noise (Amentum marks, PA charges) to adjusted numbers. It is not a name that trades on a single customer or on macro rates the way a cyclical would — it trades on mix-shift toward higher-margin, higher-growth end-markets (AI/DC) and on capital-allocation credibility.

---

Phase C — Judge people & books

Management

• CEO: Bob Pragada — CEO since Jan 2023 (~3.4 yrs), also Chair. 17 years at Jacobs (SVP Global Sales → GVP → COO Nov-2019 → CEO), with an outside stint as President/CEO of the Brock Group and COO at Kinetics (semiconductor/pharma-bio process solutions — directly relevant to the Advanced Facilities franchise). Ex-US Navy Civil Engineer Corps / Seabees officer (9 yrs).
• Track record (quantified): (1) Co-led the CH2M integration — cited internally as the largest, "highly successful" acquisition in Jacobs' history; (2) executed the Amentum Reverse Morris Trust spin cleanly (tax-free, on schedule) — genuine portfolio surgery that lifted margins and cut gov exposure to 8%; (3) drove adjusted EPS +15.9% (FY25) and +22.4% (Q2-FY26) with two consecutive guidance raises. This is a management team that has delivered structural value, not just ridden the cycle.
• Skin in the game: Pragada owns ~0.23% (~$32M) directly. Meaningful in dollars, modest as a % — professional-manager, not founder-owner, ownership. No our figures on the shelf to assess recent buy/sell behavior — a gap; a targeted refresh should ingest it.
• Capital-allocation history: Disciplined and shareholder-friendly, but just levered up. FY2025: $754.1M buybacks ($1.22B remaining auth) + a growing dividend (~1.08% yield). Then the £1.2B PA buyout (mostly cash + debt) — a high-conviction bet on the highest-margin asset it already owned, which is strategically sound (captures 100% of ~22%-margin profit) but tripled net debt to ~$2.7B and consumed the balance-sheet cushion. The judgment call: buying out a crown-jewel minority at a full price vs. buybacks/de-levering. Defensible, but it removes optionality and raises execution stakes.
• Founder vs professional manager: Classic professional manager / operator-integrator archetype — exactly the right profile for a serial-acquirer services roll-up, but it means the thesis rests on continued disciplined capital allocation, not founder vision. No red flags on comp/related-party in the filings reviewed; strategy has been coherent (focus → higher margin → AI-infra), not promotional zig-zagging.

Forensic Red Flags

Every figure labeled; sourced from the two filings + the regulatory file.

Accounting-risk map:

1. Goodwill/intangibles concentration — the biggest structural item. Goodwill $4,763–4,781M ≈ 42% of total assets ($11.9B). Legacy of CH2M + PA. No impairment taken, but a services firm carrying goodwill ~1.4x total equity ($3.29B) means any prolonged demand shock in I&AF or PA could force a write-down. Watch, not fail.
2. GAAP↔cash / non-GAAP-vs-GAAP divergence. Management steers hard to adjusted net revenue / adjusted EBITDA / adjusted EPS. The adjustments are large and recurring-ish (restructuring/transaction charges: −$162.9M FY25, −$352.2M in Q2-FY26 alone; intangible amortization −$155.5M FY25). FY2025 adjusted EPS $6.12 vs GAAP continuing $2.58 — a ~2.4x gap. Much of the gap is legitimately one-time (Amentum marks, PA-deal charges, amortization of acquired intangibles), but the sheer size and frequency of "one-time" items in a serial acquirer warrants skepticism — verify each add-back is genuinely non-recurring, not evergreen restructuring.
3. SBC. Stock-based comp $61.0M FY2025 (down from $74.2M) — modest (~0.5% of revenue), not a flatter-the-non-GAAP concern here. The PA equity-grant vesting ($207M) is a discrete event, now fully expensed ("no remaining unrecognized compensation costs… in the future").
4. Working capital / cash conversion. Receivables + contract assets rose to $3,555.6M at Q2-FY26 (from $2,989.1M); FY2025 OCF fell to $686.7M (from $1,054.7M) partly on a −$177.9M receivables build. For a services firm, watch DSO creep — it can flatter revenue recognition ahead of cash. Not alarming yet; trending the wrong way.
5. Leverage step-up. Long-term debt $2,236.5M → $4,084.2M (+$1,848M) between Sept-25 and Mar-26 to fund the PA buyout; redeemable NCI (the PA put liability) went to $0. Net debt ~$2.7B. Not distressed for a $1.2B-EBITDA business (~2.3x net-debt/adj-EBITDA ) but the cushion is gone.
6. Contingencies — the "Consolidated JV Matter." An unfavorable interim ruling against a 50%-owned consolidated JV required a reserve that dented I&AF revenue/OP in FY2025; the noncontrolling partner's share runs through NCI. Ongoing litigation/arbitration (personal injury, professional liability, breach of contract) is normal for E&C but the JV matter is a specific, unresolved item — monitor for adverse final ruling.

Regulatory findings (required sub-section):

• SEC Litigation Releases / AAERs: None. "No LR found" and "No AAER found" for Jacobs Solutions in the 2021-07-06→2026-07-06 window per SEC EDGAR EFTS.
• Non-SEC enforcement (web search): No material recent (2024–2026) DOJ/FTC/environmental enforcement surfaced. Historical only: a qui-tam False Claims Act matter alleging failure to meet employment-eligibility/drug-testing requirements on a $112M SOCOM IT contract (Jacobs Technology/Engineering) — complaint unsealed Jan-2020, the federal government declined to intervene. An older FCA matter re: rental-cost charging after a HQ sale-leaseback. FY2024: "no significant fines or non-monetary sanctions for environmental non-compliance". A Jacobs Engineering Group v. USA (Fed. Cl. 2024-cv-02053) contract dispute exists but appears to be Jacobs pursuing a claim, not a government enforcement action.
• 10-K Item 3 (Legal Proceedings): The 10-K discloses ordinary-course litigation typical of E&C (personal-injury, professional-liability, breach-of-contract) plus the Consolidated JV Matter and a non-US tax indemnity tied to the separated SpinCo business. No single item flagged as individually material to solvency.
• Verdict: No material regulatory or accounting-enforcement findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-07-06. The forensic risks are quality-of-earnings (non-GAAP add-back size, goodwill concentration, DSO creep, fresh leverage), not fraud.

---

Phase D — Project & stress-test

Forward Projection

Built bottom-up from management's own guidance + segment trajectory. Fiscal years FY2026–FY2028 (FY ends late Sept). All outputs ``; anchored to the adjusted-EPS framework the company and Street use (GAAP EPS is too mark-distorted to project cleanly).

Anchor: FY2025 adjusted EPS $6.12; FY2026 guidance $7.10–$7.35 (midpoint ~$7.22). "Challenge Accepted" targets: 6–8% organic net-revenue growth FY25–29, 16%+ adj EBITDA margin and 10%+ FCF margin by FY29.

Reasoning: Jacobs is an operating-leverage + mix-shift story, not a volume story — GAAP revenue grows mid-single-digits, but adjusted EBITDA and EPS grow low-double-digits because (a) the mix tilts toward higher-margin PA + Advanced Facilities/DC, (b) PA is now 100% owned, and (c) buybacks shrink the share count (~118M and falling). The swing factor is how durably the AI-data-center pipeline (+400%) converts to billed net revenue — the entire premium multiple rests on it.

Forecast-tracker note: Per --watchlist rules, no our model create logged in the unattended loop (only log a Brier forecast on genuine committed conviction). The base call to hand to our position log if promoted: "J FY2026 adjusted EPS ≥ $7.10 (within/above guidance), p≈0.80, resolves 2026-09-25."

Bull vs Bear

Bull case. Jacobs is the highest-quality, most liquid, asset-light way to own the AI-infrastructure buildout without owning the compute or the megawatts — it sells the scarce engineering + program-management + digital layer that gigawatt-scale AI data centers cannot be built without. The moat is a decades-deep advanced-facilities reference base that puts it on a shortlist of a handful of firms hyperscalers (NVIDIA, Hut 8, PsiQuantum) will trust with multi-billion-dollar, must-not-fail projects. The engine is already inflecting: DC revenue +100% YoY, pipeline +400%, AI-infra 10–11% of the book growing >40%, record $27B backlog (+22%), two straight guidance raises. Layered on top: a genuinely differentiated ~22%-margin PA Consulting arm now 100% owned, buybacks ($1.22B auth) shrinking the count, and a management team with a proven record of value-accretive portfolio surgery (CH2M, the clean Amentum spin). Mid-teens adjusted-EPS growth on an asset-light, ~15% ROE-ish services model deserves a premium services multiple. If the DC cycle is truly "still in early stages" (CEO), FY28 adjusted EPS ~$9–10 is achievable and today's ~17x FY26 adj P/E is undemanding on out-year numbers.

Bear case (permanent-impairment risks).

1. The multiple already prices the DC ramp. At ~14.2x EV/EBITDA (above the ~13.3x peer average) and ~23% below a $155 median PT that itself assumes the AI-infra story, there is little margin of safety — a single quarter of DC pipeline not converting, or hyperscaler capex digestion, re-rates the stock down before earnings even fall. AECOM at ~9.5x on higher margins is a live "why own the expensive one?" alternative.
2. This is still a competitive-procurement services business with ~8% core-segment margins. The 10-K itself warns competition pressures price and margin; a services roll-up is only as good as its next contract win and its labor cost. It is not a franchise with pricing power — the moat is relative, not absolute.
3. Balance-sheet cushion just spent. Net debt tripled to ~$2.7B (~2.3x adj EBITDA) to buy out PA. That removes the optionality that made J resilient, and combined with goodwill at 42% of assets and the unresolved Consolidated JV Matter, a demand shock could hit both earnings and the balance sheet (write-down) at once.

Pre-mortem (18 months out, thesis broke): It's early 2028. Hyperscaler AI-capex growth decelerated through 2027 as GPU-supply and power constraints slowed data-center starts; Jacobs' +400% pipeline converted to backlog but slipped in the schedule, so billed net revenue on DC undershot. Legacy infrastructure/water grew only low-single-digits. The premium multiple compressed toward AECOM's; adjusted EPS came in at the bear ~$7.3 instead of ~$8+, and with net debt at ~2.5x, buybacks were paused. The stock is ~$100, and the "AI-infra picks-and-shovels" narrative that justified 17x is now "cyclical E&C at 12x."

Are multiples too high? Slightly. ~14x EV/EBITDA / ~17x FY26-adj P/E is fair for the growth if it materializes, but it is a full price with no cushion — a "great company, not-great entry point" setup.

Contrarian view (what the market is refusing to see): The consensus frames J as an "AI data-center play" and pays up for it — but the durable, under-appreciated value is PA Consulting + Advanced-Facilities margin mix, which compounds regardless of any single DC cycle. Conversely, the market may be under-pricing the risk that "one-time" restructuring/transaction add-backs are semi-permanent for a serial acquirer, making the ~2.4x GAAP-to-adjusted EPS gap less pristine than the bulls' models assume.

Devil's Advocate (short-seller)

Dismantling the bull case:

• What structurally breaks the money-machine: Jacobs sells labor hours dressed up as outcomes. Strip the AI narrative and it's a low-single-digit-organic-growth, ~8%-core-margin, competitively-bid E&C firm whose GAAP earnings are a rounding error ($2.58 continuing EPS) after you stop believing the add-backs. The entire premium is the DC story — and DC is ~10% of the book. The other ~90% grows GDP-ish.
• Concentration risk: Not customer-concentrated (a positive), but thesis-concentrated on the AI-capex cycle and geography-concentrated (US 62% + Europe 24% = 86%). A US infrastructure-spending slowdown or a hyperscaler capex pause hits the growth engine directly.
• Why the moat is weaker than bulls think: It's a shortlist moat, not an exclusivity moat. AECOM, WSP, Fluor, and specialist mission-critical firms bid the same gigawatt jobs. On the very metric that matters (EV/EBITDA on superior margins), AECOM is ~33% cheaper — the market is already telling you J's premium is fragile.
• Most dangerous competitor bulls underestimate: AECOM — higher segment margins (~16.8% TTM), a cheaper multiple, and the same DC/advanced-facilities positioning. If capital rotates to "cheaper + better-margin," J de-rates.
• Worst capital-allocation move: Paying £1.2B (a full price) to buy out the PA minority with debt at the top of the cycle, tripling net debt and spending the cushion — right before a potential AI-capex digestion. If DC slows in 2027, that timing looks aggressive.
• Assumptions that must hold for today's price: (1) DC/AI-infra keeps growing >40% and converts the +400% pipeline on schedule; (2) group adjusted EBITDA margin marches toward 16% by FY29; (3) the ~2.4x adjusted-vs-GAAP add-back gap stays "one-time"; (4) no goodwill impairment / adverse JV ruling; (5) buybacks continue despite higher leverage.
• If growth disappoints 20–30%: Adjusted EPS growth halves to low-single-digits (bear ~$7.3 FY28), the premium multiple compresses toward peers/AECOM, and you can see ~20–30% downside to ~$95–100.
• Single scenario that permanently impairs: A multi-year AI-capex retrenchment that turns the +400% pipeline into slipped/cancelled projects, coincident with a goodwill write-down (42% of assets) and an adverse Consolidated-JV final ruling — hitting earnings, book value and the multiple simultaneously. Plausibility: low-to-moderate (the buildout is real and early), but non-trivial and un-cushioned.

Management Questions (ordered by information value)

1. Of the +400% data-center pipeline, what fraction is contracted backlog vs. identified opportunity, and what is your realistic conversion rate and schedule to billed net revenue over FY26–FY28?
2. How much of the ~14.6–14.9% FY26 adjusted EBITDA margin is mix (DC/PA) vs. cost action — and what pulls you to the 16%+ FY29 target if DC growth normalizes?
3. Post the £1.2B PA buyout, net debt is ~2.3x adjusted EBITDA. What is your target leverage, and does it pause buybacks or constrain M&A if the AI-capex cycle softens?
4. What is the actual operating margin of the data-center/Advanced-Facilities work vs. the I&AF segment average (~8%) — is DC accretive or just additive?
5. Your GAAP-to-adjusted EPS gap was ~2.4x in FY2025. Which of the recurring "restructuring/transaction" add-backs will still exist in FY2027, and when does the P&L stop carrying separation/PA-deal charges entirely?
6. How exposed is the DC franchise to hyperscaler capex concentration — name-level, what share of DC net revenue is your top 3 clients (NVIDIA/Hut 8/PsiQuantum-type)?
7. On the Consolidated JV Matter: range of potential final outcomes, timing, and worst-case P&L/cash impact?
8. Goodwill is 42% of assets. Under what demand scenario would I&AF or PA goodwill face an impairment test, and how much headroom exists today?
9. What is the incremental margin on the next dollar of DC/AI-infra net revenue vs. legacy infrastructure — i.e., quantify the operating leverage bulls are underwriting?
10. AECOM trades ~33% cheaper on higher margins. What do you do operationally that justifies Jacobs' premium multiple to a shareholder choosing between you?
11. Talent is the binding constraint. What is engineering-headcount wage inflation and attrition in mission-critical/DC, and can you staff the pipeline without eroding margin?
12. Capital-allocation priority stack for the next 3 years: buybacks vs. de-lever vs. tuck-in M&A vs. dividend growth — in order?
13. PA Consulting at 100% ownership: what is the standalone organic growth and margin outlook, and is there integration/dis-synergy risk now that the minority is bought out?
14. FCF margin guide is 7–8.5% vs. a 10%+ FY29 target — what specifically closes that gap, and how does DSO/working-capital (receivables rose ~$567M in H1) behave as DC scales?
15. If the AI-data-center investment cycle is "still in early stages," what would tell you first that it has turned — the leading indicator you watch internally?

---