Autodesk, Inc. (NASDAQ: ADSK) — A Toll Road on the Built World, Priced for a Detour

Independent fundamental research. Sections 1–15 below are written in a deliberately position-free, no-price-target style. The one exception is the clearly-labeled Claude's Take block immediately below, which is the author’s own subjective view.

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information only — not investment advice. The analysis that follows takes no position and names no price target.

Verdict: BUY-the-fear / ACCUMULATE-ON-WEAKNESS. Medium conviction. A live — but minority-probability — value-trap tail. Directional zone: I’d be a scaled buyer in the ~low-$180s to low-$200s (roughly EV $38–43B, ~13–15x EV/FCF, ~6–7% FCF yield on the FY27 guide), adding more aggressively toward ~$170 if AI/macro fear forces a deeper reset. This is not a short and, on the evidence, not a falling knife — it is an abandoned high-quality compounder.

Autodesk is a 91%-gross-margin, ~23%-ROIC, ~36%-ROE recurring CAD/BIM franchise that sits at a genuine regulatory-and-format chokepoint of how the physical world gets designed — and it trades at the cheapest valuation in its own ten-year history (composite own-history valuation percentile 2.4th; GAAP P/E percentile 0.09th; P/S ~1st). A reverse-DCF at $194 prices in only ~5% annual FCF growth for a decade into a business that just raised FY2027 guidance to ~+14% FCF growth and has compounded revenue ~13–14% for five years. The market is paying a value multiple for an oligopoly toll road. The framing is contrarian/value, not momentum — the factor data confirm it: negative momentum loading, relative strength ~43% off its peak, ~26% idiosyncratic volatility (the move is Autodesk-specific, not a beta wipe-out), and yet a ten-year annualized return of ~+12.8% still underneath. Three real-but-overpriced fears are doing the damage — generative-AI seat compression, the disliked $3.6B all-cash MaintainX acquisition, and decelerating “underlying” growth — stacked on the scar tissue of the 2024 accounting investigation and the loss of the Starboard catalyst (it exited in March 2025). The single most important tell that insiders see value here: the CEO, CFO, and a director bought ~$1.8M of stock on the open market in 2026 at the lows (code-P, no 10b5-1 plan) — the first conviction cluster in years.

What keeps conviction at medium rather than high: the AI-seat question is a genuine, unfalsifiable-for-now structural risk to a per-seat model; the largest acquisition in company history is landing while a 9%-of-workforce sales reorg is still settling; and the activist who would have forced the issue is gone. Bullish flip: net-new ARR re-accelerates to low-teens with visible AI-consumption (Flex-token) monetization and a clean MaintainX integration. Bearish flip: seat/ARR data confirm AI-driven per-seat erosion, or growth fades below ~6% with MaintainX dragging margins and credibility. Tag: the toll road is being priced as if the traffic is about to take a detour — I think it widens the road.

📈 Stock Price Action — Five-Year Event Map

Autodesk has round-tripped a full cycle. From a ZIRP/SaaS-mania peak close of ~$344 (24-Aug-2021) it fell ~52% to a trough of $164.31 (16-Jun-2022) in the rate shock, recovered to a fresh 52-week high of $324.42 (8-Sep-2025), and has since de-rated ~40% to $193.82 (18-Jun-2026) — printing a fresh 52-week low on 11-Jun-2026. The 52-week range is roughly $194–$324; the stock trades below its 21-, 50- and 200-day moving averages (~$221 / $231 / $257), with a beta near 1.04 and deeply negative recent alpha. Price moves are FACT; attributed causes are INTERPRETATION.

Cycle narrative. (1–2) Autodesk peaked with the 2021 SaaS bubble and de-rated violently as rates rose — a multiple event, not a business event (revenue grew throughout). (3) The 2022–2023 recovery was the subscription annuity re-rating as billings, RPO and FCF normalized. (4) Early 2024 brought the most damaging governance episode in years: an internal investigation found management had timed cash flows to hit FCF/non-GAAP targets, delaying the 10-K; Starboard filed soon after. (5) The stock nonetheless ground to its September-2025 high on a margin-expansion plan and a genuine FY26 growth re-acceleration to +17.5%. (6) From late 2025 the de-rate resumed — the whole design/creative-software complex sold off on AI-disruption fear (Adobe included), and Autodesk lost its activist catalyst when Starboard exited in March 2025. (7–8) The May-2026 MaintainX announcement — the largest acquisition in company history, all cash — was the proximate trigger for the final leg to a 52-week low, despite an accompanying earnings beat and a raised full-year outlook.

1. Executive Summary

Autodesk is the dominant design-software franchise for the people who design and build the physical world — architecture, engineering and construction (AEC), product manufacturing, and (residually) media & entertainment. FY2026 (year ended 31-Jan-2026) revenue was $7,206M, +17.5%, ~94–97% recurring, at a ~91% gross margin, generating $2,419M of free cash flow (33.6% FCF margin) and a ~23% return on invested capital. The economic core is a regulatory-and-format chokepoint: in its strongest and largest segment, AECO ($3,583M, +22%, ~50% of revenue), the proprietary Revit .rvt file format has no third-party write-back, and government BIM mandates increasingly require Revit-compatible deliverables — a switching cost measured in months of lost productivity layered on top of a regulatory ratchet. That moat shows up in the numbers: 91% gross margins, 100–110% net revenue retention, and 3–5% annual price realization.

The franchise is, however, asymmetric. The moat is deep in AECO/Revit, thinner in AutoCAD (the .dwg format is quasi-open and faces genuine low-cost competition from BricsCAD/ZWCAD), and weakest in Media & Entertainment (a $332M legacy line eroding to free tools Blender and Unreal). Manufacturing ($1,379M) is a credible mid-market challenger but a challenger, not an incumbent, against Dassault and Siemens at the high end.

The investment tension is the gap between business quality and price. After a ~40% drawdown from its September-2025 high, ADSK trades at its cheapest valuation in a decade on its own history — EV/Sales ~5.7x (vs ~12.8x in FY2022), EV/FCF ~15–17x, a ~6–7% FCF yield, and own-history valuation percentiles in the low single digits (composite 2.4th). A reverse-DCF embeds only ~5% long-term FCF growth, against management’s raised FY2027 guidance (revenue $8.155–8.215B, FCF $2.725–2.8B, ~+14%) and a five-year ~13–14% revenue CAGR.

The de-rate has identifiable, real causes, which is why this is not a no-brainer: (1) generative-AI seat-compression risk to the per-seat subscription model — the same fear that re-rated Adobe; (2) the $3.6B all-cash MaintainX acquisition (28-May-2026), the largest ever, viewed as expensive “diworsification” into maintenance/operations software and landing while a 9%-of-workforce sales reorganization is still settling; (3) decelerating underlying growth once the ~3.5pp New-Transaction-Model accounting tailwind is stripped out; and (4) residual management-credibility damage from the 2024 investigation, compounded by the loss of the Starboard catalyst. Set against those: economics that genuinely improve with scale, a remediated incentive structure, a clean balance sheet, and a 2026 insider open-market buying cluster (CEO, CFO, a director; ~$1.8M, code-P) at the lows — the first conviction signal in years.

This memo concludes that Autodesk is a structurally good business in a good industry, with a real but concentrated moat, high-quality (if magnitude-inflated) growth, adequate-not-exceptional capital allocation, and a valuation that prices a permanent impairment the evidence does not yet support — while acknowledging that the AI-seat question is a genuine, currently-unfalsifiable structural risk that justifies some of the discount.

2. Business Overview

What Autodesk sells. Autodesk makes the authoring and collaboration software used to design, engineer, build, make, and increasingly operate physical things. Customers are architects, structural/MEP and civil engineers, general and specialty contractors, product designers and manufacturers, and (in a small legacy line) film/TV/game studios. The products are mission-critical professional tools — the system of record for a building model or a manufactured part — sold on annual or multi-year subscriptions and, increasingly, on consumption (enterprise token pools and pay-as-you-go “Flex” tokens). The model is ~94–97% recurring; Autodesk has not had an annual revenue decline since the FY2017 subscription-transition trough.

Revenue by product family (FY2026, per the 10-K MD&A).

Revenue by geography (FY2026): Americas $3,178M (44%), EMEA $2,794M (39%, +21%), APAC $1,234M (17%). By type: subscription $6,743M + maintenance $33M (≈94% from those two lines; Autodesk reports ~97% recurring including the recurring portion of “other”). Forward visibility is strong: remaining performance obligations $8,300M (+20%), current RPO $5,479M (+23%), deferred revenue $4,693M (+14%), and billings $7,771M (+30%).

The platform strategy — “design → make → operate.” Historically Autodesk monetized design (CAD/BIM authoring). Over the last decade it built out make — Autodesk Construction Cloud and the “Make” product line reached ~$796M (+22%) in FY2026, anchored by PlanGrid, BuildingConnected and Forma (the cloud-native AEC platform formerly Spacemaker). The May-2026 MaintainX acquisition adds operate — computerized maintenance management (CMMS) and asset operations — forming a new “Autodesk Operations Solutions” unit. The strategic logic is a data flywheel: capture the building/asset through its full lifecycle, feed operational data back into predictive digital twins and AI. Whether that adjacency creates value or dilutes focus is the central capital-allocation debate.

How it makes money. Near-zero marginal cost of software delivery (capex is ~$33M, <0.5% of revenue) means ~91% gross margins and ~90%+ FCF conversion of operating cash flow. The business is an annuity with pricing power, not a project business — the key swing variable is net seat/subscription growth × price, both of which are macro- and (prospectively) AI-sensitive.

Verdict: A high-quality, asset-light, overwhelmingly recurring software annuity sitting at a structural chokepoint of the construction and manufacturing value chains — clearly understandable at its core, with complexity layered on by multiple product lines, an ongoing billing-model change, and heavy non-GAAP adjustment.

3. Industry Dynamics

Structure: a profitable, stable oligopoly. Global design/CAD/BIM software is controlled by four firms — Autodesk, Dassault Systèmes, Siemens Digital Industries Software, and PTC — which together command roughly 65% of the CAD market, with regionally/vertically strong specialists around the edges: Bentley Systems (infrastructure engineering), Nemetschek/ArchiCAD (European architecture), and Trimble (field/survey technology). Market structure varies sharply by vertical, which is the single most important fact about Autodesk’s competitive position:

• AEC/BIM authoring (Autodesk’s stronghold): Revit is the dominant BIM platform in North America and much of Europe; Autodesk holds an estimated 35–45%+ share of core AEC design. Competition here is getting less intense as BIM mandates standardize on Revit-compatible deliverables.
• Manufacturing (contested): Fusion/Inventor compete in the mid-market but face Dassault (CATIA/SolidWorks) and Siemens (NX/Teamcenter) at the enterprise/high end. Stable, intense, low share-shift.
• Construction field/operations (more competitive): Autodesk Construction Cloud competes with Procore, Oracle/Aconex and Trimble — a genuine, ongoing land grab.
• Media & Entertainment (structurally challenged): Maya/3ds Max face free, fast-improving Blender and real-time engines (Unreal).

Market size and growth. A realistic total addressable market for Autodesk’s current portfolio is roughly $50–65B (global CAD ~$24B, BIM/AEC ~$10–15B, construction tech ~$10B, PLM/manufacturing ~$10–15B, M&E ~$3–5B), growing ~8–10% annually. At $7.2B of revenue, penetration is ~11–14% — real runway. Management’s headline “$100B+ Design & Make” TAM, and a newly-cited “$40B operations” TAM, stretch into digital twins, operations and AI services Autodesk does not yet meaningfully sell — treat those as aspirational marketing, not investable today.

The BIM-mandate regulatory ratchet. The strongest secular tailwind is the global spread of government BIM mandates: the UK (BIM Level 2 on public projects since 2016, >70% professional adoption), Germany (federal infrastructure), Italy, Singapore, Malaysia, with India developing national policy. These are not reversible cyclical trends — they are regulatory ratchets that permanently expand Autodesk’s addressable base and, crucially, often require Revit-format deliverables, converting technical lock-in into regulatory lock-in. Layer on the data-center construction boom (a fast-growing AEC vertical, ~+30% YoY) and infrastructure spending.

Capital-cycle read (Marathon). This is a high-return industry that is not attracting destabilizing new capacity — the format moats, regulatory standardization, 10–20-year platform-development lead times, and the trained-installed-base network effect deter entry. The last meaningful new core-CAD entrant, Onshape (2015), was absorbed by PTC for $470M. The one force that could lower the barrier to entry — and is the genuine cycle risk — is generative AI shrinking the cost and time to build credible design tools.

Verdict: a structurally GOOD industry — oligopolistic, high-margin (peers run 75–91% gross / 25–48% operating margins), high barriers to entry, a regulatory tailwind in the largest segment — with one real structural wildcard (AI) that the next five years will adjudicate.

4. Competitive Position

The moat is real, financially proven, and concentrated. Name it in Greenwald’s taxonomy: demand-side captivity through switching costs, reinforced by an intangible (proprietary file-format + industry-standard) advantage, a regulatory lock-in, and education-funnel / ecosystem network effects. The financial fingerprint confirms it: ~91% gross margin (stable for six years), ROIC ~23% (up from ~16% in FY2022 and steady FY2023–26), net revenue retention 100–110%, and 3–5% annual price realization with limited churn. These outcomes would deteriorate without the moat — pricing power and retention are the moat made visible — so it qualifies as a moat under our test (a “moat” that cannot be tied to a financial outcome is not one; this one can).

But the moat is asymmetric — pressure-test it segment by segment:

• Strongest — AECO/Revit. The .rvt format is proprietary with no third-party write-back: rivals (BricsCAD, others) can import .rvt as static 3D solids but cannot export an editable .rvt. Leave Revit and you abandon your parametric BIM model — years of families, templates, standards and multi-discipline coordination. Industry participants consistently describe migration as 6–12 months of disrupted operations. Stack on BIM-mandate deliverable requirements and the moat here is deep and arguably deepening.
• Thinner — AutoCAD/.dwg. The .dwg format is quasi-open; BricsCAD and ZWCAD natively read and write .dwg at materially lower prices, and have won cost-driven switchers (e.g., THS Concepts, PERI). Lock-in here is habit/ecosystem, not format — genuine low-end price competition exists. AutoCAD still grew +14% in FY2026, so the captivity is real, just shallower.
• Weakest — M&E. A $332M legacy line whose +5% is partly contra-revenue accounting rather than demand; structural erosion to free Blender and Unreal. Immaterial to the thesis (4.6% of revenue) but an honest tell that Autodesk’s moat is not uniform.
• Manufacturing. Fusion wins SMB/mid-market on price and cloud-native workflow, but Autodesk is a challenger, not the incumbent, against Dassault and Siemens in aerospace/automotive PLM.

Versus key competitors. Autodesk’s 91% gross margin tops the design-software group (Bentley ~80%, PTC ~84%, Dassault ~85%, Nemetschek ~75%), the signature of near-zero delivery cost and pricing power. Its ~38% non-GAAP operating margin, however, trails PTC’s ~48% on similar gross margins — the gap Starboard hammered and the clearest evidence that the cost structure carries slack the franchise economics should not require.

Verdict: a durable but concentrated advantage. Autodesk passes the competitive-advantage test clearly in AECO/Revit (where ~half of revenue and the most durable moat sit), less convincingly in AutoCAD and Manufacturing, and fails it in M&E. The consolidated ~23% ROIC and 100–110% NRR confirm the franchise is real; the asymmetry is why the bull case rests disproportionately on AEC.

5. Growth History and Forward Opportunities

History: real, organic, decelerating — and magnitude-inflated by accounting. Revenue compounded from $3,274M (FY2020) to $7,206M (FY2026), a five-year CAGR of ~13.7%, overwhelmingly organic (~85–95%; the largest deal, Innovyze ~$1B, added perhaps 2–3 points in FY2022). The subscription transition (2016–2020) — which cratered revenue ~19% in FY2017 before tripling it — is the prior chapter; the current one is steady mid-teens growth with margin expansion.

The critical nuance: FY2026’s headline +17.5% overstates underlying growth by ~3.5 points. The New Transaction Model (NTM) — the shift from a reseller buy-sell model to direct agency billing — grosses up reported revenue (partner margin now flows through Autodesk’s top line). Management confirmed the NTM added ~3.5pp to revenue growth and ~1.5pp to billings in Q1 FY2027, fading to ~2pp in Q2 and ~1.5pp for the full year. Strip it and Autodesk is a low-teens grower decelerating toward ~10% — still high quality (recurring, 91%-margin, price-led), but the optics flatter the trajectory by 2–3 points for a couple of years before normalizing.

Forward drivers (credible, execution-dependent):

• AECO / BIM secular expansion — the strongest engine: expanding global BIM mandates, the data-center construction boom, and the “Make”/Construction Cloud business at ~$796M (+22%).
• Pricing power — 3–5% annual increases plus the hidden price increase of eliminating reseller discounting under the NTM (AEC trade press estimated resellers were giving away ~$600M).
• Mix-shift upsell — individual products → AEC Collections (~$3,225/yr vs ~$1,975 for AutoCAD) → enterprise token pools (EBAs, ~15% of revenue) and consumption (Flex).
• AI monetization — Neural CAD (generative 3D producing editable BREP geometry), Autodesk Assistant (agentic AI across Fusion/Revit/AutoCAD), Forma; the consumption-pricing model is the value-capture vehicle.
• Operations adjacency — MaintainX/AOS as a new growth vector (and TAM-expansion claim).
• Residual conversion — remaining perpetual/non-compliant users, a diminishing but not exhausted pool.

The AI seat-compression debate — the single most important forward variable. Bear: if generative AI automates 80–90% of routine design tasks (Autodesk’s own Neural CAD claim), the logical endgame is fewer designer seats per project, directly threatening a model where ~94%+ of revenue is per-seat subscription — the same fear that re-rated Adobe. Bull/management: Autodesk owns three scarce assets generic models can’t replicate — proprietary CAD/BIM geometry (BREP) training data, workflow context, and deep 3D engineering expertise (~100 peer-reviewed papers) — and runs a hybrid architecture (“AI generates; deterministic parametric/physics engines validate”) for a domain where “probably right” is unacceptable (buildings must be correct). Crucially, Autodesk already has a consumption-pricing rail (Flex tokens) to capture value per AI-use even if seats compress — a structural hedge most peers lack. On balance AI more likely strengthens than destroys the franchise, but the consumption transition is unproven at scale and the direction-of-travel risk is genuine.

Verdict: high-quality growth, overstated magnitude. Recurring, organic, high-margin, pricing-led — but the headline rate is ~3.5pp of accounting tailwind over a low-teens-and-decelerating underlying core, and the long-run rate hinges on the unresolved AI-monetization question.

6. Financial Quality

Revenue, margins, and the GAAP/non-GAAP gap. Revenue $7,206M (+17.5%) at a ~91% gross margin. Here the numbers require care: GAAP operating margin was 22% (income from operations $1,578M) — not the ~25% some data aggregators report (a mis-grouping of operating expense; reconcile to the 10-K). Non-GAAP operating margin was 38% ($2,737M). GAAP diluted EPS $5.23; non-GAAP diluted EPS $10.43 (non-GAAP net income $2,242M). The ~$5.20 GAAP-to-non-GAAP EPS gap is large and recurring, and ~68% of it is stock-based compensation ($788M, ~11% of revenue), with ~2% from intangible amortization and ~3% from restructuring. SBC is a real, recurring economic cost — it consumed ~33% of FY2026 FCF — and non-GAAP metrics exclude it; we weight GAAP and owner-economics accordingly.

Cash flow and the billing-model distortion. FCF: FY2024 $1,283M → FY2025 $1,545M → FY2026 $2,419M (+57%), a 33.6% FCF margin; Q1 FY2027 FCF $876M (+58%). The FY2024 trough and FY2026 surge are not a clean signal — they are the unwind and normalization of the multi-year-upfront → annual billing transition (and, in the 2022–2024 window, of management’s now-disclosed cash-timing decisions; ). The FY2027 FCF guide of $2,725–2,800M is the better run-rate read. Quality of earnings is otherwise high: GAAP net income ~$1.1B against ~$2.5B operating cash flow is the normal subscription pattern (cash collected ahead of ratable recognition), and there are no signs of accrual-vs-cash deterioration.

Returns and balance sheet. ROIC ~23%, ROE ~36%. ROE is flattered by a buyback-shrunk, negative-tangible-book equity base (cumulative repurchases and the transition-era deficit), so ROIC is the honest read — and ~23% on an invested-capital base that includes ~$4.3B of acquisition goodwill comfortably clears a ~10% WACC. The balance sheet is clean: cash + short-term investments ~$2.92B against total debt ~$2.73B at 30-Apr-2026 — essentially net cash pre-MaintainX, with interest coverage >20x. (Pro forma for the $3.6B MaintainX deal, Autodesk moves to modest net debt, <1x EBITDA.)

Verdict: economics genuinely improve with scale — gross margin stable at the top of the peer set, operating leverage real (incremental operating margins ~38–39%), FCF conversion high, capex negligible — but the reported picture is distorted upward by (a) a ~3.5pp NTM revenue gross-up, (b) a large SBC-driven GAAP/non-GAAP wedge, and © billing-transition cash-flow noise. Normalize all three and you still have a very good business, just a less spectacular one than the headlines suggest.

7. Capital Allocation

The buyback is value-neutral, not value-creating — the central criticism. Autodesk pays no dividend and returns cash almost entirely via repurchases — ~$1.40B in FY2026 (5M shares). Yet over five years, despite ~$4.5B+ of cumulative buybacks, diluted shares fell only ~222M → ~212M (−4.5%). The arithmetic is unflattering: most of the spend simply offsets SBC dilution (~2–3%/yr). This is capital allocation that prevents value destruction by compensation rather than creating per-share value — the single most legitimate critique of the franchise, and the strongest part of the (now-departed) Starboard thesis: cut SBC as a share of revenue and the per-share math improves materially.

R&D and operating discipline. R&D was $1,643M (~22.8% of revenue), appropriately high for a platform business and trending down as a percentage — a source of the margin expansion. Management targets ~41% non-GAAP operating margin by FY2029 (from 38% in FY2026), i.e., ~300bps of further expansion.

M&A track record — mixed, and now a big new bet. History: BuildingConnected ($275M) and Spacemaker/Forma ($240M) look astute; PlanGrid ($875M) anchored construction cloud at a high price; Innovyze ($1B, 2021) is the cautionary one (Autodesk paid ~4x what EQT had paid four years earlier). The new and defining transaction is MaintainX — $3.6B all-cash, announced 28-May-2026, the largest in company history, at ~27x forward ARR (>$135M ARR growing >50%). It is financed by cash plus a new $1.0B 364-day delayed-draw term loan and an upsized revolver ($1.5B → $2.0B), excluded from FY2027 guidance, expected to close in FY2027. The market’s objection is threefold: a rich multiple, a move away from core design into operations/CMMS, and the timing — integrating the largest deal ever while a 9% sales reorganization is still settling. Management’s defense is the “construction playbook” (a cornerstone acquisition plus tuck-ins, citing the ~$1.8B construction investment that became ~$600M of >20%-growth revenue) and the data flywheel into predictive operations/AI. This is the swing capital-allocation risk of the thesis — plausible strategically, expensive and execution-stacked in practice.

Compensation and incentive alignment — materially improved. CEO Andrew Anagnost’s FY2025 total compensation was ~$30.1M (pay ratio ~203.5:1) — high, and a fair governance criticism. But the incentive design has been remediated in direct response to the 2024 episode: the annual plan is Revenue (60%) / non-GAAP income-from-operations (40%), and — importantly — the long-term PSU metric was changed to “non-GAAP income from operations less SBC,” and free cash flow was removed as a metric. Removing the FCF target that management had been caught steering, and charging SBC against the incentive pool, is a genuine fix to the exact incentive that produced the problem.

Insider behavior — a 2026 conviction buy cluster. The signal worth weighting: in 2026, at the lows, named insiders bought stock on the open market (code P, no 10b5-1 plan) — CEO Anagnost ~$499K @ $202.66, CFO Moorjani ~$494K @ $197.67, and Director Stacy Smith ~$794K @ ~$231 (~$1.8M combined), following Chairman John Cahill’s ~$534K purchase during the March-2025 proxy fight. Against a 2025 backdrop of routine sell-into-strength (10b5-1) at $310–325, this is a clear contrarian tell that those closest to the numbers see value at current prices.

Verdict: adequate, not exceptional, and improving. The clean balance sheet, disciplined R&D, and remediated incentives are positives; the value-neutral buyback and the expensive, focus-stretching MaintainX bet are the offsets. Management has not yet earned an “intelligent capital allocator” verdict — but the incentive fix and insider buying argue against the “empire-builder” caricature.

8. Changes and Headwinds — Last Two Years

1. The 2024 audit-committee investigation (the credibility scar). An internal investigation (disclosed in the late-filed FY2024 10-K, June 2024) found that during FY2022–FY2024 management “relied on multi-year contracts billed upfront to help meet free cash flow targets,” ran early-renewal/upfront-billing incentive programs, and made “decisions regarding discretionary spending, collections, and accounts payable … informed by their anticipated effects on … external free cash flow and/or non-GAAP operating margin targets.” No restatement was required; the SEC and the U.S. Attorney (SDNY) closed their inquiries without action by ~August 2025. CFO Deborah Clifford was replaced by Janesh Moorjani (ex-Elastic). Residual effect: investors must apply some discount to management’s forward FCF/non-GAAP guidance — trust, once dented, rebuilds slowly — but the incentive remediation directly addresses the root cause.

2. Starboard Value — campaign and exit. Starboard built a ~$500M stake (13D, Feb-2024) and pushed for board change and a 45% non-GAAP operating margin by FY2028. Autodesk added two directors (Jeff Epstein, Christie Simons). Then Starboard exited its entire position in early March 2025 — ending the campaign and removing the activist catalyst. The governance improvements are lasting; the pressure that would have forced faster margin action is gone (a double-edged development).

3. Workforce reduction / GTM reorganization. In February 2025 Autodesk cut ~1,350 employees (~9% of workforce) and restructured its go-to-market organization (partly to fund the margin plan and AI investment). Management flagged elevated near-term disruption risk — a real, ongoing execution headwind into FY2027.

4. New Transaction Model rollout — the agency/direct-billing shift (84%+ migrated by FY2025), which improves pricing control and data ownership but grosses up reported revenue and created channel friction.

5. MaintainX — the $3.6B all-cash push into operations (28-May-2026).

6. AI product cadence — Neural CAD, Autodesk Assistant, Forma, and the $200M World Labs strategic investment.

7. Macro. US construction spending fell ~4.7% in real terms in 2025 and tariff policy clouds manufacturing capex — partly offset by the data-center construction boom (~+30%).

Verdict: net neutral-to-mildly-positive for business quality, but net negative for near-term execution risk. The governance overhang is largely remediated and the activist-era discipline modernized the company; simultaneously, the largest-ever acquisition, a 9% reorg, and a soft construction backdrop are stacked into the same 12–18 months. The thesis improves structurally and gets riskier tactically at the same time.

9. Risk Analysis

Catastrophic-loss risk: very low. Near-zero net debt (pre-deal), ~$2.4B+ annual FCF, ~94%+ recurring revenue, and workflow-embedded products make mass defection structurally implausible; even in an extreme AI-disruption scenario the installed base would be a multi-tens-of-billions strategic asset long before equity impairment. The realistic downside is de-rating and growth disappointment, not impairment.

10. Valuation Discussion (Embedded Expectations)

Where it trades. At $193.82 (18-Jun-2026) on ~211M shares, market cap is ~$40.9B; with ~net cash, EV ≈ $40.7B. That is:

Against peers, Autodesk now sits at or below the cheap end of design software despite a top-of-group 91% gross margin and ~23% ROIC: ADBE EV/Sales ~7–8x (also AI-de-rated), Dassault ~7–8x, Trimble ~5–6x, PTC ~9x, Bentley ~11–12x, Nemetschek ~11x; quality anchors Cadence/Synopsys ~15–18x and Intuit ~9–10x. Only severely-re-rated Dassault is clearly cheaper.

Embedded-expectations reverse-DCF. Discounting Autodesk’s FCF at ~10% with a ~3% terminal growth rate, the EV of ~$40.7B prices in only ~4.8% annual FCF growth off the FY2026 $2.42B base — and ~3.0% off the FY2027 guided ~$2.76B base. That is the crux: the market is underwriting ~5% perpetual FCF growth for a 91%-gross-margin, ~23%-ROIC franchise that just raised FY2027 FCF guidance to ~+14% and has compounded revenue ~13–14% for five years. The bar is set near GDP-plus.

Scenarios (EV):

• Bear (~$34–39B): AI seat compression bites, growth stalls to mid-single digits, FCF plateaus ~$2.5–3B, multiple stays ~12–14x EV/FCF. Roughly flat-to-down from here — this is the value-trap path.
• Base (~$55–63B): ~8–10% revenue growth, margins drift toward the 41% target, FCF ~$3.3–3.6B by ~FY2029 at ~16–18x — a re-rate toward (still below historical) multiples on a larger FCF base.
• Bull (~$75B+): growth re-accelerates with visible AI/consumption monetization and clean MaintainX execution, FCF ~$3.5–4B at ~20x.

What the market is pricing correctly: genuine growth deceleration as the NTM tailwind fades; the real (if unquantifiable) AI-seat risk; MaintainX integration/cash-out and focus risk; and a credibility discount. What it appears to be pricing incorrectly (interpretation): treating a high-margin, regulatory-moated annuity as a ~5%-grower in perpetuity, and a high-margin tuck-in as value-destructive before any evidence, while the same company guides FCF up ~14% and insiders buy. No price target — this section sizes expectations, not a target.

11. Variant Perception

Consensus. The sell side is structurally bullish — roughly two-dozen analysts skew Buy/Strong-Buy with essentially no Sell ratings — but price targets were cut after the MaintainX deal to ~$305–330 (from ~$340–370). So consensus holds the direction (good business, undervalued) while marking down the magnitude on deal/AI worries. The stock trades well below even the cut targets.

Strongest bull case. A regulatory-and-format-moated oligopoly toll road at the cheapest valuation in its own history, where a reverse-DCF demands only ~5% FCF growth against ~14% guided; a remediated incentive structure; a clean balance sheet; an under-appreciated consumption-pricing rail (Flex) that lets Autodesk monetize AI rather than be disrupted by it; and an insider buy cluster. Re-rate to even ~25x FCF on $3.5B (~FY2029) implies an EV near ~$87B — substantial upside from ~$41B.

Strongest bear case. Generative AI structurally compresses designer seats faster than the consumption model can offset; “underlying” growth is already fading to low-teens and headed lower; MaintainX is expensive empire-building that dilutes margins and focus; management’s credibility is impaired and the activist who would have enforced discipline is gone; and a 2.4th-percentile multiple can persist for years (the value-trap path) absent a catalyst. In that world FCF stagnates ~$2.5–3B and the multiple stays ~13–15x — flat-to-down.

The 3–5 assumptions that matter most: (1) whether AI is net-accretive or net-dilutive to seat economics over five years — the fulcrum; (2) the durability of ~8–10% organic growth as the NTM tailwind fades; (3) MaintainX accretion vs. distraction; (4) sustaining ~33% FCF conversion through the integration and reorg; (5) whether the multiple re-rates at all without an activist or an ARR re-acceleration. Falsification — bull: net-new ARR/seat growth re-accelerates to low-teens with visible AI-consumption revenue and clean MaintainX integration. Falsification — bear: seat/ARR disclosure confirms AI-driven per-seat erosion, or organic growth falls below ~6% with MaintainX dragging margins.

Factor-positioning read (the tape as evidence). The empirical positioning supports contrarian/value, not falling knife. Autodesk shows a negative momentum loading once sector/industry factors are stripped, relative strength ~43% off its peak, deeply negative recent alpha, and it trades below all major moving averages — yet its idiosyncratic volatility is high (~26%), meaning the drawdown is an Autodesk-specific story, not a beta wipe-out, and a strong recent Quality-factor tailwind did not lift it (the market is trading it as a busted single-name, not re-rating it as the quality compounder it screens as). Underneath sits a ~+12.8% ten-year annualized track record. This is the factor signature of an abandoned high-quality name — consensus may be offsides on the magnitude of permanent impairment it is pricing — while honestly flagging that the value-trap tail is live precisely because the catalyst (Starboard) has departed.

12. Fact vs. Interpretation

13. Open Questions

1. AI seat economics: Will AI features net-grow or net-shrink paid seats over five years, and how fast can Flex/consumption monetization offset any seat compression? (Autodesk does not yet disclose seat counts or consumption revenue cleanly.)
2. Underlying growth: What is the ex-NTM, ex-price organic seat/volume growth rate — and is it stabilizing or still decelerating?
3. MaintainX economics: What are MaintainX’s gross margins, net retention, and CAC, and what is the integration/margin path? Does AOS become a third growth pillar or a distraction?
4. Margin bridge to 41%: How much of the FY2029 non-GAAP op-margin target depends on SBC reduction vs. genuine operating leverage, and will SBC as a % of revenue actually fall?
5. Buyback intensity post-MaintainX: Does the ~50%-of-FCF buyback survive the deal’s cash/debt draw, and will net share count finally decline meaningfully?
6. Construction cycle: How does ~94% recurring revenue behave through a genuine multi-year AEC downturn — never tested under the current subscription model?

14. What Must Be True

For the bull case (re-rating toward fair value):

• The AEC/Revit moat holds and deepens (BIM mandates, construction-cloud adoption, AI integration), keeping NRR ≥100% and pricing power intact.
• Organic growth stabilizes ~8–10% as the NTM tailwind fades — not mid-single-digits — and AI generates incremental consumption revenue rather than cannibalizing seats.
• Margins march toward the ~41% FY2029 target with SBC declining as a share of revenue, lifting FCF toward ~$3.5B+ and finally shrinking the share count.
• MaintainX integrates cleanly and the operations adjacency is at least margin-neutral.
• Falsification test: if, over the next 2–4 quarters, net-new ARR/seat metrics fail to stabilize in the high-single-to-low-teens and AI-consumption revenue remains immaterial, the re-rating thesis is wrong.

For the bear case (value trap / impairment):

• Generative AI compresses per-project seat counts faster than consumption pricing can offset, pushing organic growth below ~6%.
• MaintainX dilutes margins and focus; the FY2029 margin target slips; FCF stagnates ~$2.5–3B.
• The credibility discount and absent catalyst keep the multiple at a 2.4th-percentile level for years.
• Falsification test: if Autodesk sustains ~$2.7B+ FCF growing ~10%+, holds ~38%+ non-GAAP margins through the reorg/integration, and net share count finally falls, the value-trap thesis is wrong.

15. Source Appendix

Key sources: Autodesk FY2022–FY2026 Forms 10-K (esp. FY2026, filed 03-Mar-2026; FY2024, filed 10-Jun-2024 with the audit-committee Explanatory Note); Q1 FY2027 Form 10-Q (filed 29-May-2026) and the 28-May-2026 earnings release/8-K; the 28-May-2026 MaintainX acquisition announcement and the June-2026 financing 8-Ks; DEF 14A proxies (compensation/incentive metrics, board) and DFAN14A (Starboard); Forms 3/4/5 (insider transactions); the Q1 FY2027 and Q4 FY2026 earnings-call transcripts; aggregated fundamentals/ratios and public market-price/valuation data; factor/risk-model data; and public industry/trade sources as cited.

APPENDIX A — Standard Diligence Questionnaire — Autodesk, Inc. (NASDAQ: ADSK)

As-of 2026-06-19. Supplemental to the research memo. Fact / Interpretation / Assumption labels applied where material.

General

What thoughtful questions have other investors asked about this company? The sharpest investor questions cluster on four themes. (1) NTM durability — how much of FY2026’s +17.5% growth is the one-time New-Transaction-Model accounting gross-up (~3.5pp) versus sustainable organic expansion (the honest answer: underlying growth is low-teens and decelerating). (2) The margin gap — why does Autodesk run a ~38% non-GAAP operating margin when PTC runs ~48% on similar gross margins (Starboard’s core critique). (3) AI seat compression — if Autodesk’s own Neural CAD automates 80–90% of routine design tasks, doesn’t the endgame mean fewer seats? (4) Credibility — after the 2024 investigation revealed management timing cash flows to hit metrics, how much should investors discount forward FCF/non-GAAP guidance? Interpretation.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Tilting to the favorable end, not a peak. Fact: FY2026’s +17.5% was the fastest growth since the subscription transition, but ~3.5pp was NTM accounting and the rest rode a favorable AECO/data-center backdrop; FY2027 guidance decelerates to ~+12–13% headline. Interpretation: earnings quality is high (94%+ recurring) but the growth rate is temporarily elevated by the billing model and will normalize.

Driven by external environment or internal actions? Predominantly internal — the subscription transition, the NTM repricing, 3–5%/yr price increases, and Collections/EBA upsell are management-driven, layered on real-but-secondary external tailwinds (BIM mandates, data-center construction). New seat growth is the macro-sensitive piece.

How stable are revenues? Extremely, by software standards: ~94–97% recurring, 100–110% net revenue retention, no annual revenue decline since FY2017. The risk is growth deceleration, not revenue collapse — an outright decline would require mass non-renewals that Revit/BIM switching costs make unlikely.

Outlook for products/services? Strong in AECO (BIM mandates, construction cloud), adequate in Manufacturing (mid-market wins, enterprise challenger), structurally challenged in M&E (Blender/Unreal erosion; 4.6% of revenue, immaterial). MaintainX/operations is a new, unproven adjacency.

How big will the market be? Realistic current-portfolio TAM ~$50–65B growing ~8–10%; Autodesk’s $7.2B is ~11–14% penetration — real runway. Geographic split ~44% Americas / 39% EMEA / 17% APAC, fastest growth in APAC/emerging markets. Assumption: management’s “$100B+” and “$40B operations” TAMs are aspirational; treat $50–65B as investable.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Less in AEC/BIM authoring (mandates standardize on Revit), stable-intense in Manufacturing CAD (same four players 20+ years), more in construction field software (Procore/Oracle/Trimble) and much more in M&E (free Blender). The largest/strongest segment is becoming more defensible — what matters most.

How profitable — ROIC/ROE? Fact: ~91% gross margin, ~23% ROIC (FY2026), ~36% ROE (flattered by negative-tangible-book, buyback-shrunk equity — ROIC is the honest read), 100–110% NRR. Clears a ~10% WACC comfortably but is “very good,” not “monopoly-grade,” partly because ~$4.3B of acquisition goodwill and ~11%-of-revenue SBC dilute per-share economics.

How profitable is the industry / barriers to entry? One of tech’s most profitable — peers at 75–91% gross / 25–48% operating margins. Barriers stack: 10–20-year platform development, proprietary file formats, trained-installed-base network effects, regulatory (BIM-mandate) lock-in, and data gravity. The last meaningful core-CAD entrant (Onshape, 2015) was absorbed by PTC.

Can the business be easily understood? Yes at the core (sticky subscription software for designers who can’t easily switch); complexity comes from multiple product lines, the billing-model change, and heavy non-GAAP adjustment.

Undermined by foreign low-cost labor? No — Autodesk sells tools, not labor. Offshoring of architects/engineers expands its market (more global BIM adopters).

Do brands matter? Yes — “AutoCAD” and “Revit” are near-generic terms and default project specifications; the education funnel (free student licenses) reinforces it. But switching costs and format lock-in, not brand, do the moat’s heavy lifting.

Nature of competition? Capability/ecosystem/workflow entrenchment rather than price (BricsCAD in 2D CAD the exception); more vertical/geographic market-sharing than head-to-head price war — the hallmark of a healthy oligopoly.

Customer switching costs? Very high for Revit/BIM (6–12 months of disrupted operations; no third-party .rvt write-back; contractual Revit-deliverable requirements), moderate for AutoCAD (.dwg is quasi-open; BricsCAD reads/writes it), low for M&E.

Financial Condition & Balance Sheet

Assets not on the balance sheet? Substantially — the ~7M+ recurring-subscriber base, the AutoCAD/Revit brands, and decades of proprietary CAD/BIM training data (the AI moat) are all off-balance-sheet. These are why the business is worth ~$41B against deeply negative tangible book.

Off-balance-sheet liabilities? Operating leases (~$400–500M), cloud-infrastructure purchase obligations, and future SBC dilution (~2–3%/yr before buybacks). No material contingencies; the SEC/DOJ matters closed without action.

How conservative is the accounting? Moderately aggressive. Revenue recognition is straightforward (ratable) with clean EY opinions, but the 2024 investigation showed management steering operational cash decisions to hit non-GAAP/FCF targets, and the non-GAAP framework excludes a large recurring SBC charge (non-GAAP EPS $10.43 vs GAAP $5.23 in FY2026). Presented in the most favorable defensible light.

How CapEx-hungry? Minimal — capex ~$33M (<0.5% of revenue); OCF-to-FCF conversion ~90%+. A signature attraction.

Capital Allocation & Management

How much FCF, and how used? FY2026 FCF $2,419M (33.6% margin); FY2027 guide $2,725–2,800M. Allocated almost entirely to buybacks (~50% of FCF policy), no dividend. The criticism: ~$4.5B+ of five-year repurchases reduced diluted shares only ~222M → ~212M (−4.5%) — the buyback mostly offsets SBC dilution rather than creating per-share value.

Significant acquisitions recently? Yes — MaintainX, $3.6B all-cash (28-May-2026), the largest ever, ~27x ARR, into maintenance/operations (CMMS). Prior: Innovyze ($1B, 2021, expensive), PlanGrid ($875M), BuildingConnected ($275M), Spacemaker/Forma ($240M), World Labs ($200M strategic investment).

Buying back shares? Yes (~$1.40B/5M shares FY2026) — but net-of-SBC the share count barely moves. Anti-dilution, not value creation.

Issuing large amounts of stock to insiders? Yes — SBC $788M (~11% of revenue, ~33% of FCF), in line with large-cap software peers (Adobe/Salesforce/Intuit) but the biggest single drag on per-share value.

Compensation policy? CEO Anagnost ~$30.1M FY2025 (pay ratio ~203.5:1) — high. Materially improved incentive design: annual plan = Revenue 60% / non-GAAP income-from-operations 40%; long-term PSU metric changed to “non-GAAP IfO less SBC,” and FCF removed as a metric — a direct remedy to the incentive that produced the 2024 episode.

Motivations of management? Equity-comp-driven (now better aligned after the metric change); a meaningful but not outsized founder-CEO-class stake. The 2026 open-market insider buying (CEO ~$499K, CFO ~$494K, Director Smith ~$794K; code P) is a genuine alignment/conviction signal at the lows.

Valuation & Market Data

ADR / MLP / K-1? No — a domestic C-corp on NASDAQ, standard 1099 reporting, no K-1.

Dividend policy? None, no stated intention. (At a ~6–7% FCF yield, a dividend might be more shareholder-friendly than the value-neutral buyback — but management shows no interest.)

Net income diverging from cash from operations? Yes, in the healthy direction — GAAP NI ~$1.1B vs OCF ~$2.5B, the normal subscription pattern (cash collected ahead of ratable recognition), with the FY2024 FCF dip explained by the billing transition and since normalized.

Risks & Downside

What would cause the stock to decline? (1) Evidence of AI-driven seat compression; (2) FY2027 execution miss from the 9% reorg + MaintainX integration; (3) a sharper construction downturn breaking the “secular growth” narrative; (4) further credibility erosion; (5) a broader software de-rate. Interpretation: the most probable decline path is growth disappointment + the multiple staying compressed (value trap), not a fundamental break.

Risk of catastrophic loss? Very low — near-zero net debt (pre-deal), ~$2.4B+ FCF, ~94%+ recurring revenue, workflow-embedded products.

Chance of total loss? Effectively zero — even in a severe AI-disruption scenario the installed base and FCF make Autodesk a multi-tens-of-billions strategic asset long before equity approaches zero.

Recent News & Events

Has the business environment changed recently? Yes — the $3.6B MaintainX acquisition (28-May-2026) is the defining recent event (push into operations, market disliked it); AI competitive dynamics are evolving fast (Neural CAD, $200M World Labs); US construction spending fell ~4.7% real in 2025 (offset by data-center construction ~+30%); tariff policy clouds manufacturing capex.

Significant acquisitions? MaintainX (above) — the largest in company history.

Accounting-policy changes? Enhanced internal controls after the 2024 investigation (no restatement); the NTM shifted agency-model revenue recognition (net→gross), inflating reported growth ~3.5pp; new CFO emphasizes more conservative communication.

Recent business changes / new management? (1) ~1,350-employee (~9%) layoff + GTM reorg (Feb-2025); (2) new CFO Janesh Moorjani (ex-Elastic); (3) two new directors (Jeff Epstein, Christie Simons) from the Starboard settlement, then Starboard’s full exit (Mar-2025); (4) Autodesk Construction Cloud/“Make” at ~$796M (+22%); (5) Neural CAD / Autodesk Assistant AI launches; (6) FY2027 guidance raised above prior consensus.

APPENDIX B — Source Appendix — Autodesk, Inc. (NASDAQ: ADSK)

Report date 2026-06-19. Primary sources prioritized. All URLs accessed 2026-06-19 unless noted. Fact / Interpretation distinctions are made in the memo body.

Primary — SEC filings

1. Form 10-K, FY2026 (period ended 31-Jan-2026; filed 03-Mar-2026) — revenue by product family/geo/type, segment growth, gross/operating margins, GAAP→non-GAAP reconciliation, RPO/deferred revenue, SBC, buybacks, risk factors, competition. https://www.sec.gov/Archives/edgar/data/769397/000076939726000015/adsk-20260131.htm
2. Form 10-K, FY2025 (filed 06-Mar-2025). https://www.sec.gov/Archives/edgar/data/769397/000076939725000013/q425pressrelease.htm (release); 10-K under CIK 0000769397.
3. Form 10-K, FY2024 (filed 10-Jun-2024 — late, with the Audit-Committee Explanatory Note disclosing the FCF/non-GAAP/billing-timing investigation). https://www.sec.gov/Archives/edgar/data/769397/000076939724000101/ (CIK index)
4. Form 10-K, FY2022 & FY2023 — multi-year revenue/margin/FCF history.
5. Form 10-Q, Q1 FY2027 (period ended 30-Apr-2026; filed 29-May-2026) — Q1 results, billings/RPO, NTM tailwind quantification, balance sheet. https://www.sec.gov/Archives/edgar/data/769397/000076939726000044/adsk-20260430.htm
6. Form 8-K / earnings release, Q1 FY2027 (28-May-2026) — Q1 revenue $1,934M, non-GAAP EPS $2.99, GAAP/non-GAAP op margins 28%/39%, FCF $876M; FY2027 guidance raised (rev $8.155–8.215B, billings $8.505–8.58B, FCF $2.725–2.8B). https://www.sec.gov/Archives/edgar/data/0000769397/000076939726000041/q127pressrelease.htm
7. 8-K — MaintainX acquisition (28-May-2026) — $3.6B all-cash agreement. https://www.sec.gov/Archives/edgar/data/769397/000121390026062125/ea029248301ex99-1.htm
8. 8-Ks — June 2026 (filed 15-Jun-2026 and 17/18-Jun-2026) — acquisition financing (new $1.0B 364-day delayed-draw term loan; revolver upsized to $2.0B).
9. DEF 14A proxy statements (FY2024–FY2026) — CEO/NEO compensation (~$30.1M FY2025; pay ratio ~203.5:1), EIP/PSU incentive metrics (Revenue 60% / non-GAAP IfO 40%; PSU = “non-GAAP IfO less SBC”; FCF removed), board composition.
10. DFAN14A / DEFA14A (Mar-2025 and 2024–2025) — Starboard Value proxy-contest materials (filer agent CIK 0000921895) and Autodesk responses.
11. Forms 3/4/5 — insider transactions, incl. 2026 open-market purchases (code P): CEO Anagnost (~$499K @ $202.66), CFO Moorjani (~$494K @ $197.67), Director Stacy Smith (~$794K @ ~$231); Chairman John Cahill (~$534K, Mar-2025); 2025 routine 10b5-1 sales at $310–325.

Primary — Earnings-call transcripts

12. Q1 FY2027 earnings call (28-May-2026) — MaintainX rationale, AI strategy (Neural CAD / Autodesk Assistant / hybrid generate-and-validate), NTM tailwind (~3.5pp rev / ~1.5pp billings), margin and guidance commentary.
13. Q4 FY2026 earnings call (Feb-2026) — FY2026 results, FY2027 framework, margin-target commentary.
14. FY2023–FY2024 earnings-call transcripts (context for the subscription/billing transition and the 2024 investigation period).

Quantitative data (reconciled to filings)

15. Aggregated fundamentals data — income statement, balance sheet, cash flow, profitability ratios (ROIC ~23%, ROE ~36%), enterprise value, valuation multiples (multi-year). Note: one aggregator’s FY2026 GAAP operating income/margin was mis-grouped (~$1,794M/24.9%); corrected to the 10-K figure of $1,578M / 22%.
16. Public market price history — daily split/dividend-adjusted OHLCV, EMAs, beta/alpha. 5yr high ~$344 (24-Aug-2021); trough $164.31 (16-Jun-2022); 52-wk high $324.42 (08-Sep-2025); close $193.82 (18-Jun-2026).
17. Own-history valuation percentiles: P/E 0.09th, P/S ~1st, P/B 6th, composite 2.4th.
18. Public analyst-action coverage — (Wells Fargo PT to $330, RBC to $305, both 29-May-2026; MaintainX-deal scrutiny).
19. Factor/risk-model data — factor loadings (negative momentum once sector/industry stripped), leaderboard (rs_peak ~−43; idiosyncratic vol ~26%; ~+12.8% 10-yr annualized), related-stocks comp cross-check. (Endpoints dated 2026-06-17.)

Secondary — trade press, deal coverage, industry

20. Bloomberg, “Autodesk to Buy MaintainX for About $3.6 Billion in Cash” (28-May-2026). https://www.bloomberg.com/news/articles/2026-05-28/autodesk-to-buy-maintainx-for-about-3-6-billion-in-cash-mppxfnfb
21. PR Newswire / Autodesk News, “Autodesk to acquire MaintainX, advancing unified platform in operations” (28-May-2026). https://www.prnewswire.com/news-releases/autodesk-to-acquire-maintainx-advancing-unified-platform-in-operations-302784938.html
22. AEC Magazine, “Autodesk buys MaintainX for $3.6bn” (May-2026). https://aecmag.com/operations/autodesk-buys-maintainx-for-3-6bn/
23. SiliconANGLE, “Autodesk to acquire MaintainX for $3.6 billion in push into operations” (28-May-2026). https://siliconangle.com/2026/05/28/autodesk-acquire-maintainx-3-6-billion-push-operations/
24. Invezz / TradingView, “Autodesk stock falls as $3.6B MaintainX deal worries investors” (29-May-2026). https://invezz.com/news/2026/05/29/autodesk-stock-falls-as-3-6b-maintainx-deal-worries-investors/
25. StockTitan, “Autodesk Q1 FY27 revenue up 18%, FY27 view raised” (28-May-2026). https://www.stocktitan.net/news/ADSK/
26. Benzinga, “Autodesk Gains Analyst Backing, But MaintainX Deal Faces Scrutiny” (29-May-2026). https://www.benzinga.com/analyst-stock-ratings/analyst-color/26/05/52881847/
27. Autodesk News, “Autodesk issues statement in response to comments by Starboard Value” (26-Mar-2025) — Starboard’s 45%-margin-by-FY2028 demand and Autodesk’s margin/FCF rebuttal. https://adsknews.autodesk.com/en/pressrelease/autodesk-issues-statement-in-response-to-comments-by-starboard-value/
28. Boardroom Alpha, “Starboard Challenges Autodesk: Proxy Fight Looms Over Governance & Performance” (2025).
29. BricsCAD Help Center / AEC Magazine (2025) — .rvt import-only (no third-party write-back) and .dwg read/write interoperability; BricsCAD/ZWCAD competitive pricing.