Autodesk, Inc.

Conclusion: Autodesk’s economic engine is still strengthening: it is a mission-critical design software tollbooth with high switching costs, increasingly recurring revenue, and little evidence that the core franchise is fading.

Autodesk (ticker: ADSK; trading currency: USD) makes money by selling design and engineering software used across architecture, construction, manufacturing, and media. The DNA is simple: own the workflow, then monetize it through subscription seats, enterprise agreements, and cloud-linked tools. Products like AutoCAD, Revit, Inventor, Fusion, Maya, and construction/cloud collaboration software are deeply embedded in how customers design, document, and coordinate work.

The model is better today than it was a decade ago. Legacy maintenance revenue is becoming irrelevant, while subscription revenue dominates. In the six months ended July 31, 2025, subscription revenue was 3396000000? Wait no total revenue was 3396000000 and subscription revenue was 3190000000? Let's be precise: subscription revenue was 3190000000, up from 2738000000; maintenance fell to 17000000 from 22000000; total revenue rose to 3396000000 from 2922000000. That is exactly what you want to see: old revenue streams shrink while the core recurring engine grows faster.

This is mostly a win-win model. Autodesk’s software is expensive in isolation, but cheap relative to labor, rework, and project costs. If a design platform saves even small amounts of engineering time or reduces construction errors, the ROI is obvious. The caveat: customers will tolerate price increases only as long as Autodesk keeps delivering productivity and interoperability rather than just monetizing lock-in.

What would worry me? A real slowdown in subscription growth, weakening deferred/recontracted demand, or credible product displacement by cloud-native/AI-native alternatives. Today, the bigger story is not obsolescence; it is entrenchment.

If I tracked only a few numbers, they would be: subscription revenue growth, total revenue growth, deferred/recontracted revenue indicators, and operating/free-cash-flow margins. Those tell you whether Autodesk is still deepening its grip on the workflow or merely harvesting a mature install base.

Autodesk’s market is a long-term tailwind: it sits at the center of the global “design-and-make” software stack, where digitization, cloud collaboration, BIM adoption, simulation, and AI all expand software spend per user rather than commoditize it. Using company filings through January 31, 2026, the key point is that Autodesk is no longer just selling CAD seats; it is widening into connected workflows across AEC, manufacturing, and media.

The good news is structural: end markets are cyclical, but software penetration inside those markets is still rising. The bad news is also structural: competition is real, especially wherever workflows become more open, collaborative, and cloud-native.

Conclusion: Autodesk has a real, durable moat, driven mainly by switching costs, workflow embeddedness, and ecosystem scale—not by patents or pure brand. Using the most recent annual data (FY2026 ended January 31, 2026), the moat looks stable to modestly strengthening, but not dramatically widening.

The proof is in the revenue model: in the July 2025 half-year, subscription and maintenance were 3,207,000,000 dollars of 3,396,000,000 dollars revenue, and deferred revenue remained very large at 3,844,000,000 dollars. That is what entrenched software looks like. The moat is not a toll bridge or network effect; it is a deeply embedded system-of-work moat.

Autodesk’s financial strength is strong: this is a high-return, cash-rich subscription software model with manageable debt, though stock-based compensation and buyback-distorted equity make headline earnings and ROE less clean than the business economics.

Most recent annual data used: FY2026 ended January 31, 2026.

Conclusion: Autodesk still has a long reinvestment runway, but the best opportunities are mostly internal product, platform, and workflow expansion rather than balance-sheet-heavy M&A. As of July 31, 2025, this remains an asset-light compounding model: subscription revenue grew to $3,190,000,000 in the first six months of FY2027, while capex was just $17,000,000. That spread is the tell. Incremental capital is going into R&D, cloud delivery, industry workflows, and go-to-market - and because much of that is expensed, GAAP ROIC likely understates the true economic return.

The implied organic growth rate looks high single digit to low double digit, supported by price, seat expansion, construction cloud, and deeper wallet share in AEC/manufacturing. The main caveat is that retained earnings are not the binding constraint; Autodesk already generates more cash than it needs for maintenance reinvestment, so excess cash will often flow to buybacks rather than new high-ROIC projects.

Buybacks have created value mainly because the core business still compounds; M&A has been selective, not serial. Return on incremental invested capital remains strong.

Autodesk is a category leader, not the whole category leader. In AEC/BIM and general-purpose design authoring, it is the franchise to beat; in heavy industrial PLM and high-end manufacturing, Dassault, Siemens Digital Industries Software, and PTC are tougher. The most relevant public peers are PTC and Bentley in the U.S., and Dassault Systemes globally; Siemens is strategically important but less comparable as a segment inside a larger group.

Market share looks slightly up where Autodesk matters most: Revit remains the BIM standard, AutoCAD is still the default drafting layer, and Construction Cloud/Fusion expand wallet share. The company is not clearly winning enterprise PLM against Dassault/PTC/Siemens, but it does not need to; its AEC base is larger, stickier, and more standardized. The long-term outlook is for share gains at the workflow level, with pricing, cloud attach, and cross-sell now more important than pure seat growth.

Conclusion: reasonably aligned, but not exemplary. Autodesk behaves more like a shareholder-conscious software compounder than an empire-builder, yet management credibility is not spotless.

Net: aligned enough to own, not aligned enough to trust blindly. Latest financial filing used: FY2026 10-K, year ended January 31, 2026.

Conclusion: Autodesk’s management looks good, not spotless: capital allocation has mostly been sensible and execution has been strong, but the 2024 reporting-controls episode prevents an “elite stewardship” verdict.

They have largely allocated capital toward the right things: product investment, tuck-in M&A, and meaningful buybacks. In the six months ended July 31, 2025, Autodesk repurchased $712,000,000 of stock while keeping liquidity solid ($2,003,000,000 cash; $2,481,000,000 long-term notes). Execution has been credible: recurring revenue keeps compounding and margins remain healthy. The weak spot is transparency/process discipline. Management did disclose the internal investigation and filing delay rather than bury it, but strong operators do not let KPI/reporting practices become a governance issue in the first place. I do not see a disclosed restatement or auditor disagreement in the recent 10-Ks, and current litigation appears ordinary rather than existential. Net: capable managers, but trust is not perfect.

Conclusion: Autodesk looks modestly undervalued, not dramatically cheap. At the assumed USD 44.31B market cap, the stock is being priced more like a decent software company than a dominant, high-switching-cost design platform with recurring revenue and rising cash conversion.

My preferred framework here is normalized free-cash-flow value, not liquidation value. On the user-supplied TTM FCF of about USD 3.16B, Autodesk trades at roughly 14x FCF and about 5.7x TTM revenue. For a mission-critical software franchise with 90%+ gross margin and subscription revenue, that is closer to fair-to-cheap than expensive.

Management’s near-term guidance has been credible. In its January 22, 2026 8-K, Autodesk said Q4 and full-year FY2026 billings, revenue, non-GAAP operating margin, EPS, and FCF would all finish above the top end of prior guidance. That matters: this is not a team missing numbers and asking for patience. The restructuring is real, but it looks more like margin cleanup than demand distress.

A reasonable base case is low-double-digit revenue growth plus some operating leverage as sales optimization finishes. That can plausibly produce USD 4.0B+ FCF by 2031. Put a still-reasonable 15x-16x FCF multiple on that and equity value lands around USD 60B-65B. My current intrinsic value estimate is about USD 52B-56B today, or roughly USD 245-265/share.

The market is implicitly assuming either slower growth, lower durability, or a permanent governance/forecasting discount. I think that is somewhat too pessimistic, but not absurdly so.

Hard liquidation is a bad way to own this: tangible book is negative, and after debt, shareholders would get little in a forced wind-down. The value is overwhelmingly in the living franchise, not the assets.

Conclusion: Autodesk still looks ownable for a 5-10 year holder, but this is more likely a solid compounder than a spectacular one; roughly 2-3x in 10 years is plausible if the moat stays intact.

The moat should hold a long time because Autodesk sits inside design workflows, file standards, trained labor pools, and compliance-heavy project ecosystems. That usually erodes slowly. What breaks first is not core CAD demand; it is the ability to keep bundling more value into the platform faster than cloud-native, lower-cost, or AI-assisted alternatives can unbundle it.

Reinvestment still works. Incremental R&D behind Fusion, construction cloud, data interoperability, and AI can deepen workflow lock-in and raise wallet share. But returns on new capital are unlikely to stay exceptional forever: the core AEC franchise is already mature, so incremental returns should drift from excellent toward merely good rather than collapse.

Even under adverse conditions, Autodesk should still matter in 10-20 years. The failure case is not disappearance; it is becoming a slower-growth utility.

Thesis-break signal: customers stop accepting price increases and subscription growth decelerates to mid-single digits or worse for several periods while renewal quality, deferred revenue, or cross-sell into construction/manufacturing weakens. That would mean the moat is narrowing, not just the cycle softening.

Conclusion: Autodesk’s risk profile is manageable, not trivial. The core franchise is durable; the real bear case is not macro softness but technological displacement of the design workflow itself.

The only risk that could permanently impair Autodesk is a slow loss of control over mission-critical authoring standards if AI-native, cloud-native, or more interoperable competitors make AutoCAD/Revit less central to customer workflows. If that happens, pricing power and renewal durability would erode together. Probability: low-to-moderate, roughly 1-in-4 over a decade; impact: very high.

Other issues matter less. Construction/manufacturing cyclicality, longer deal cycles, FX, and regional weakness are mostly uncertainties: they can slow growth, but they do not by themselves break switching costs. Financial risk is modest; recurring revenue, high gross margins, and manageable leverage make a balance-sheet failure unlikely. Governance risk is also limited: filings do not suggest fraud or related-party abuse, though Autodesk remains exposed to the usual large-cap software risks around M&A integration, SBC optics, and execution on newer cloud products. Cybersecurity and data-sovereignty regulation are watch items, but only become true thesis-breakers if they meaningfully damage customer trust or product usability.

TRACK — great business, not a great enough price

Autodesk is an exceptional software franchise: sticky workflows, real pricing power, high returns, and a long runway in design-to-build digitization. The problem is simpler than the business: at roughly USD 44.31 billion, the stock does not offer enough upside versus your own base-case value of USD 60.00 billion by 2031. That is respectable, not compelling.

This is not a fraud, not a melting ice cube, and not a value trap. It is a high-quality compounder priced around fair value. Permanent capital-loss risk looks below average because the moat is real, cash generation is strong, and leverage is manageable. But below-average risk alone is not enough if expected returns are only middling.

The inversion case is clear: if AI-native design tools, cloud collaboration layers, or owner/operator workflow platforms weaken Autodesk’s control over the daily system of work, the company could remain profitable while losing its ability to compound at premium rates. In that case, today’s valuation leaves limited protection.

So the answer is: ownable, but not urgent. For new money, I would wait for a better entry or better evidence of durable acceleration. For existing holders, this still looks like a hold, not a sell; the business quality is too high to exit casually, but the setup is not strong enough to press aggressively. This is small-tranche territory on weakness, not “load the truck.”

Is the analysis accurate and complete? Mostly, but I would still verify:

• Whether FY2026 free cash flow benefited from working-capital timing that will reverse.
• The pace of net revenue retention and seat expansion in AEC versus manufacturing.
• How much real product advantage Autodesk has in AI workflows versus simply bundling features.
• Whether SBC plus buybacks are truly shrinking owner dilution over a full cycle.