The BIM Monopolist's $45 Chasm — Is SBC an Implicit Tax or a Growth Investment?

Autodesk (NASDAQ: ADSK) In-Depth Stock Research Report

Analysis Date: 2026-03-26 · Data as of: FY2026 (January 31, 2025)

Chapter 1: Executive Summary

One-Sentence Conclusion

Autodesk is the de facto monopolist in BIM (Building Information Modeling—a construction design technology that replaces traditional 2D drawings with 3D digital models) design software (Revit 63.5% market share), structurally protected by BIM mandates in 20+ countries, with organic growth of 12-13% sustainable for 2-3 years—but the $788M/year SBC "implicit tax" creates a $45 chasm between the standard FCF valuation ($238) and the true shareholder return valuation ($193). The core divergence lies in whether investors choose a Standard or Owner framework to price ADSK, and whether SBC can continuously converge—which itself heavily depends on revenue growth (70% of the decline in SBC/Rev comes from denominator growth)—will determine which framework ultimately prevails.

Key Figures at a Glance

Rating: Watch (+1~+10%, Lower Bound)

Rating Logic:

• Adjusted PW Standard: $238/share (+1.1%)—almost precisely fairly priced after deviation adjustment
• PE Expansion Option: 19x→22x probability 40% = +$37/share (+16%)→probability-weighted +6.4%
• Bearish Drag: 8 Bear arguments probability-weighted -$49→probability-weighted -20.8%
• Combined Expected Return: +1.1% + 6.4%×40% - 20.8%×(diversified ~15%) ≈ +5~8%
• At the boundary between "Neutral Watch (-10~+10%)" and "Watch (+10~+30%)"
• Boundary Handling: RSI (Relative Strength Index, measuring the intensity of recent buying or selling of a stock, range 0-100, below 30 is oversold) 12.45 extremely oversold + 4 PPDA (Probability-Price Divergence Analysis, used in this report to compare the gap between the "probability-weighted fair price for each scenario" and the "current market price"; a larger gap indicates potential mispricing) divergences (3 indicating undervaluation) + BIM mandate structural floor→leans towards the lower bound of "Watch"
• Difference from "Neutral Watch": If only static valuation (+1%) is considered, it should be Neutral Watch. However, ADSK is in a combination of technical extremes (RSI 12.45) + fundamental improvement (FY2026 +18%, FCF recovery)—historically, such combinations have yielded a median return of +25-45% after 6 months. We assign higher weight to technical/sentiment factors than to pure valuation, thus maintaining "Watch" rather than downgrading to "Neutral Watch".

Reversal Signal Monitoring Checklist

Chapter 2: Key Controversies: What the Market is Debating

Controversy 1: Is ADSK a "Cheap SaaS" or "Fairly Priced Masked by SBC"?

Bull Case: Forward Non-GAAP PE of 19x is the lowest in the SaaS sector (PTC 18.5x, NOW 35x, DDOG 49x). ADSK's profitability (OPM 38%) and cash conversion (FCF 33%) are superior to most peers—19x is too cheap.

Bear Case: Non-GAAP PE ignores the $788M/year SBC. Owner PE of ~37x is not cheap. If the market starts pricing with Owner Economics (CRM experienced this process in 2022-23), PE will not expand—it may contract.

Our Assessment: Both sides are half right. The standard PE of 19x is indeed low (caused by SEC discount + SaaS de-rating), but the Owner PE of 37x indicates that ADSK is not as "cheap" as it appears. The true valuation lies between the two—contingent on whether SBC converges. If SBC declines from 10.9% to 8% (FY2030), Standard and Owner PE will converge at 22-25x; if it does not converge, a Standard PE of 19x would be the fair price.

Controversy 2: Is Government-Mandated BIM a "Moat" or a "Growth Ceiling" for Autodesk?

Background: Over 20 countries globally have passed regulations requiring public construction projects to use BIM technology for design (i.e., "BIM mandate"). Autodesk's architectural design business (i.e., the AEC segment—Architecture, Engineering, Construction) contributes approximately 50% of the company's revenue, and its core product Revit is currently the BIM software with the highest market share (approximately 63.5%).

Bullish View: Government legislation mandating BIM adoption = "Legal floor" for demand in Autodesk's architecture business. This demand will not fluctuate with economic cycles (government projects must use BIM regardless of economic conditions), providing extremely high certainty.

Bearish View: The government mandate is "must use BIM technology," not "must use Autodesk's Revit"—this is a critical distinction. Other BIM software solutions are available on the market (such as Graphisoft's ArchiCAD and Nemetschek's Allplan) that can also meet compliance requirements. More importantly, if governments globally require the adoption of the open IFC standard (a universal BIM data format not tied to any vendor) in the future, rather than Revit's proprietary format, then competitors would also benefit equally. Furthermore, over 20 countries have already implemented BIM mandates, with future additions primarily being developing countries like China and India—where Revit's market share is well below 60%, making the benefit uncertain.

Our Judgment: The BIM mandate is "likely positive (80%)" for Autodesk, rather than "definitely positive." In mature markets where mandates have been implemented (UK, Germany, Singapore), Revit's market share exceeds 60%, with clear benefits; however, in newly mandated markets (China, India), Revit's market share is less than 40%, leading to uncertain benefits. Overall, the BIM mandate has been downgraded from a "definite floor" to a "high-probability floor."

Controversy 3: Is AI "Enhancement" or "Cannibalization"?

ADSK-AI Dichotomy: AI in AECO segment = enhancement (generative design + BIM data flywheel); AI in AutoCAD segment = cannibalization (automated drafting → reduced seats). Probability-weighted net AI score +1.0 (slightly positive) — but this figure is highly uncertain (only 0.5/5 of AI deep assessment invariants passed). AI is not a catalyst for ADSK — it is "uncertain low noise."

Chapter 3: Most Critical Drivers and Key Risk Signals

Variable 1: SBC Convergence Rate (Impacts $50/share = 21% of valuation)

Every 1 percentage point drop in SBC/Revenue from 10.9% → Owner FCF margin +0.8 percentage points → Owner valuation +$8-10/share. If FY2030 reaches 8.5% (-2.4 percentage points) → Owner valuation +$20-25 → The gap between Standard and Owner valuation narrows from $45 to $20. This is the most critical variable that could upgrade "Monitor" to "Deep Dive" — but note: 70% of the decline in SBC/Revenue comes from revenue growth (denominator effect) rather than absolute SBC reduction. Therefore, SBC convergence fundamentally relies on revenue growth maintaining 12%+. SBC convergence alone is not enough; it also requires P/E expansion.

Variable 2: FY2027 Q1 Performance (Catalyst Event, May 2026)

If Q1 beats + guidance is not lowered → RSI reversal confirmed + tariff fears subside → stock price could return to $260-280 (+10-20%). If Q1 misses or guidance is lowered → RSI could fall to <10 → bottoming out at $215 (52W low).

Top 5 Key Risk Signals

"Boiling Frog" Warning

The most dangerous scenario is not a single event shock, but rather KS-1 (slow growth deceleration) + KS-4 (SBC stagnation) + KS-8 (MFG lagging PTC) simultaneously and slowly deteriorating — each step appears "acceptable," but cumulatively, the P/E could compress from 19x to 12-14x over 5 years. Probability 15-20%.

Six-Method Cross-Validation

Convergence Range: $220-$300 (excluding extremes), Central Tendency $238 (Standard)/$258 (Comparable Median)
Key Observation: The $44 difference between Standard DCF ($238) and Bear Probability-Weighted ($194) ≈ SBC Capitalization Cost ($45) → ADSK's Bull/Bear case is entirely determined by the SBC narrative

Quality Scorecard

B-Record (Business Quality)

C-Record (Competitive Moat)

Overall Score

Compounding Path Classification: B-Grade (Medium Compounding) — ADSK is not A-Grade (Sustainable Compounding) because C-Record is weak (network effects/data flywheel/cost barriers are all low); nor is it C-Grade (Value Trap) because switching costs + institutional embeddedness provide a solid baseline. ADSK is a "medium-compounding enterprise living off switching costs and institutional embeddedness" — not a buy-and-hold-for-50-years company, but also not a company that will be disrupted tomorrow.

Chapter 4: Reverse Valuation: What Belief Does a $239 Price Imply?

4.1 Methodology and Parameters

Reverse Discounted Cash Flow (Reverse DCF—deriving market-implied growth assumptions from the current stock price, rather than directly calculating "how much it's worth") serves as the narrative anchor for this report. Here, we have built a complete reverse valuation model using Python, with the core parameters as follows:

Why was WACC=10% chosen as the baseline? Autodesk's Beta = 1.467. Calculated with a risk-free rate of 4.5% (10-year US Treasury bond) and an Equity Risk Premium (ERP) of 4.5%: Ke (Cost of Equity) = 4.5% + 1.467 × 4.5% = 11.1%. Because Autodesk's net debt is only $485M (extremely low leverage), WACC ≈ Ke ≈ 11%. However, ADSK's Beta has decreased from 1.6 to 1.47 in recent years because the maturity of its subscription model (97% recurring revenue) has reduced earnings volatility. Considering that Beta is still converging, we use 10% as a neutral baseline, with ±1% for sensitivity analysis.

4.2 Standard FCF Implied Growth Rate

Core Question: How fast does Autodesk's revenue need to grow over the next 5 years to justify a $239/share price?

Standard Reverse DCF Sensitivity Matrix (Implied 5-year Revenue CAGR)

Baseline scenario (WACC=10%, Terminal FCF Margin=35%) highlighted

Under the baseline scenario (WACC=10%, Terminal FCF Margin=35%, g=3%), a $239 price implies a 5-year revenue CAGR of 10.9%.

What does this number mean? It needs to be compared against three reference points:

Therefore, $239 prices in "moderate pessimism"—the market-implied growth rate is 1.6 percentage points lower than management's guidance and is close to the lower bound of Autodesk's organic growth over the past 4 years. This is neither extreme fear (which would imply <8%) nor fair pricing (which should be ≈12%), but rather a "distrust premium" lying between the two.

Where does this distrust come from? There are three quantifiable sources:

(1) Residual discount from SEC investigation. The 2024 Audit Committee investigation found that management manipulated the timing of achieving FCF/Non-GAAP margin targets. Although both the SEC and DOJ have closed their cases (on August 19 and 21, 2025, respectively) with no financial restatement, CFO Deborah Clifford was replaced by the Chief Strategy Officer. Because the core of the SEC investigation was the credibility of FCF data—and reverse DCF relies on FCF assumptions—the market's trust in Autodesk management's FCF guidance may have been permanently damaged by 2-3%.

(2) Billing transition noise. FY2026 reported revenue growth was +18%, but this included a "catch-up effect" from multi-year contracts shifting from upfront payment to annual payment—the billings growth of +30% significantly exceeding revenue growth is evidence of this. Therefore, the market might be applying a simple but rational discount: Reported growth - catch-up effect = true organic growth ≈12-13%, then subtract the uncertainty of "when the catch-up effect will truly end" → resulting in pricing at 10.9%.

(3) Restructuring Uncertainty. FY2026 restructuring expense is $216M (2,350 layoffs, approximately 16% of total employees). Management states this is a "structural GTM transformation" (shifting from renewal-based sales to new customer acquisition + AI/Cloud), but the market might interpret 16% layoffs as "business under pressure, forced to cut costs" – these two narratives correspond to completely different valuations.

4.3 Owner Economics: SBC is an Implicit Tax

Standard Reverse DCF uses reported FCF ($2.41B) as the cash flow base. However, Autodesk's SBC is $788M annually, equivalent to 32.7% of FCF. This means: for every $100 of FCF, $33 is "paid" to employees as compensation through the dilution of existing shareholders' equity – it never leaves the company as cash, but it undeniably transfers value.

Owner FCF (True Owner FCF) = FCF - SBC × (1-Tax Rate)

Re-performing the Reverse DCF based on Owner Economics, the implied 5-year revenue CAGR surges to 18.4% – which far exceeds any reasonable expectation (management guidance 12.5%, organic historical 10-12%).

This creates a sharp divergence in interpretation:

Therefore, whether Autodesk is undervalued depends on which cash flow definition you believe. If you believe SBC will eventually converge (management guidance FY2027 SBC/Rev<10%, and continues to converge towards 5-7%), then the standard FCF definition is closer to reality – in this case, $239 is undervalued. If you believe SBC is a permanent dilution cost (over $700M annually), then the Owner FCF definition is more accurate – in this case, $239 is fair or even overvalued.

This is a philosophical disagreement, not an arithmetic one. The DDOG v2.0 report observed the same phenomenon (Non-GAAP P/E 49x vs Owner P/FCF 188x) – the "fair valuation" of SaaS companies essentially revolves around the debate of "whether SBC is a cost."

4.4 Decoding Four Implied Market Beliefs

Behind the $239/share price, the market implicitly holds four specific beliefs:

Belief 1: Revenue growth will gradually decline from 18% to 10-11%

The Reverse DCF implies a 5-year CAGR of 10.9%, meaning the market expects FY2026's +18% growth (including catch-up from billing model transition) will gradually decrease to ~10% by FY2031. This aligns with analyst consensus (FY2027 +13%→FY2028 +10%→FY2029 +10%). The market is not pricing in an AI explosion (which would require >14%), nor is it pricing in a growth collapse (which would imply <8%).

Belief 2: FCF Margin will stabilize at 33-35%

Current FCF Margin is 33.4%, and FY2027 guidance implies 33-34% ($2.75B/$8.14B). The base case scenario uses a 35% terminal FCF Margin, meaning the market expects FCF Margin to have only 100-150bps of expansion room. Given that Non-GAAP OPM is already 38%, and the gap between FCF and Non-GAAP OI (primarily CapEx of $43M and Working Capital changes) is small, this assumption is reasonable.

Belief 3: Terminal Value Multiple ≈ Mature Software (14-16x FCF)

Terminal Value (TV—representing all value beyond the model's forecast period) accounts for 75.2% of EV. In the base case scenario, the implied terminal FCF multiple is ≈14.3x (=Terminal FCF/Terminal EV). This is the valuation level of mature software companies (vs. 25-35x for high-growth SaaS), suggesting the market positions Autodesk as "mature" rather than "growth" – despite its revenue still growing at +12%.

Belief 4: SBC will slowly converge but not disappear

SBC/Rev decreased from 12.6% in FY2022 to 10.9% in FY2026, with FY2027 guidance <10%. The market's implied assumption is that SBC will decline from 10.9% to ~8-9% over the next 5 years (rather than decreasing to 7.9% like PTC). Because Autodesk's PSU (Performance Share Unit) conditions are tied to revenue + FCF targets, as long as management continues to pursue growth, SBC will not significantly decrease.

4.5 WACC Sensitivity: The Parameter That Determines Everything

Reverse DCF is highly sensitive to WACC. A ±1% change in WACC alters the implied CAGR by 3.5 percentage points:

If the WACC should actually be 11% (based on the current Beta of 1.467 and a calculated Ke of 11.1%), then the implied CAGR for $239 is 14.4% – higher than guidance, meaning the market is actually pricing in "neutral to optimistic" rather than "moderately pessimistic."

This is a critical methodological caveat: our "moderately pessimistic" conclusion is based on the assumption of WACC=10%. If WACC is higher (closer to the theoretically calculated 11%), the market's pricing is actually reasonable. Therefore:

• If you believe Autodesk's Beta is still converging (from 1.6→1.2) → WACC should be lower (9-10%) → $239 is undervalued
• If you believe Beta stabilizes at 1.4-1.5 → WACC≈11% → $239 is fairly valued
• WACC determines the "undervalued vs. fair value" judgment more significantly than the growth assumption

4.6 Chapter Conclusion

The pricing signal for $239 is ambiguous, rather than clear.

Calculated using standard FCF, the implied growth expectation for $239 (10.9%) is below management guidance (12.5%), making it appear undervalued; but calculated using Owner FCF after deducting SBC, to support $239 requires 18.4% compound growth – this requirement is extremely demanding. The two definitions yield contradictory conclusions, indicating that $239 is not a "clearly undervalued" or "clearly overvalued" price.

Under what conditions is it worth buying? At least one of the following must be observed: (a) SBC as a percentage of revenue rapidly declines below 8% (improving true shareholder returns); (b) After excluding billing model transition factors, organic growth consistently exceeds 12%; (c) After the SEC investigation is resolved, market confidence recovers (requiring 2-3 consecutive quarters of clean results); (d) Stock price volatility continuously decreases and Beta converges below 1.3.

What is the biggest bearish argument? Owner FCF Yield is only 3.4% – even if the standard P/E (19x) appears below the industry average (22x), the true shareholder return after deducting SBC is not attractive. In other words, if SBC cannot significantly converge, the "cheapness" of the P/E is merely an illusion.

Chapter 5: Two-Dimensional Valuation Benchmarking

5.1 Why Two-Dimensional Benchmarking is Needed

Most analyses categorize Autodesk as "SaaS" and benchmark it against Salesforce/Adobe/ServiceNow, or as "CAD/PLM" and benchmark it against PTC/Dassault/Siemens. However, Autodesk spans two worlds—it is the only software company with #1 or #2 products in all three vertical sectors: AEC (Architecture, Engineering & Construction), MFG (Manufacturing), and M&E (Media & Entertainment). Therefore, single-dimensional benchmarking systematically misses key information:

• The SaaS dimension tells us "how the software market prices growth and profit margins"
• The CAD/PLM dimension tells us "how industry competition constrains growth and profit margins"
• The intersection of these two is ADSK's "reasonable P/E range"

5.2 SaaS Dimension: Strongest Profitability but Second Lowest Valuation

SaaS Cohort: Forward Non-GAAP P/E vs Non-GAAP OPM (Latest FY)

Data Source: FMP API + Latest Financial Reports from Each Company

Three Key Observations:

(1) ADSK is an anomaly with the "third-highest OPM but second-lowest P/E." Its Non-GAAP OPM of 38% is surpassed only by ADBE (47%) and PTC (43%), and is higher than NOW (31%) and CRM (33%). However, its Forward P/E (19x) is at the bottom, alongside CRM (16x). Since high profitability usually implies a higher valuation (higher margins → better cash flow quality → market willingness to pay a higher multiple), ADSK's low P/E represents an anomaly that requires explanation.

(2) Clear organic growth stratification explains most P/E differences. DDOG (+28%) → 49x, NOW (+22%) → 25x, ADSK/PTC/ADBE (+12%) → 18-19x, CRM (+9%) → 16x. Growth is the primary driver of P/E. Within the 12% growth bracket, 18-19x appears to be the "industry fair level." This means ADSK's P/E is not an anomaly—it is simply categorized within the "12% growth companies" group.

(3) However, ADSK's FCF Margin (33%) is the highest among the 12% growth group. ADBE 38%, PTC 31%, ADSK 33%—ADSK's cash generation efficiency is better than PTC's but lower than ADBE's. If FCF Margin were the second factor for valuation (to differentiate among companies with similar growth rates), then ADSK should command a higher P/E than PTC (>19x) but lower than ADBE (≤18x). The current 19x equals PTC's level, implying that the market is not assigning any premium to ADSK's FCF advantage.

Why no premium? Revisiting the three discount factors from Chapter 4: residual SEC investigation, billing transition noise, and restructuring uncertainty. These three factors have pulled ADSK from "should be ≥PTC (19-22x)" down to "=PTC lower bound (19x)."

5.3 CAD/PLM Dimension: Reasonable Valuation but Differentiated Competition

CAD/PLM Cohort: Valuation and Business Benchmarking

*Includes catch-up/acquisition effects   †Includes $216M restructuring   ‡Estimated based on EV/EBITDA 30.7x   Data Source: FMP + Company Financial Reports

Key Findings from the CAD Dimension:

(1) ADSK has the lowest EV/Sales among CAD/PLM peers. 7.1x vs PTC 9.3x / BSY 8.8x / Dassault 8.5x / Nemetschek 14.5x. This is consistent with the SaaS dimension's conclusion—ADSK is undervalued relative to its profitability and growth.

(2) Bentley's valuation premium (EV/EBITDA 30.7x vs ADSK 29.7x) is worth interpreting. Bentley's revenue is only $1.5B (1/5 of ADSK's), with lower growth (+11% vs ADSK +18%), and similar GAAP OPM (24% vs 22%). However, Bentley has two characteristics that ADSK lacks: extremely low SBC/Rev (4.8% vs 10.9%)—cleaner Owner Economics; and a leading digital twin (iTwin) platform—clearer AI narrative. Therefore, Bentley's premium likely stems from "more credible true profitability + a purer AI narrative."

(3) PTC's GAAP OPM of 35.9% significantly exceeds ADSK's 21.9%. However, this gap primarily results from ADSK's $216M restructuring charges and $139M acquisition amortization. If ADSK is adjusted to "excluding restructuring + amortization" = GAAP OPM ≈ (1,578+216+139)/7,206 = 26.8%—it is still 9pp lower than PTC. The gap stems from: PTC's S&M/Rev of 29.0% vs ADSK's 32.9%—PTC has higher sales efficiency (more large PLM contracts, fewer small CAD orders). This implies that ADSK indeed lags PTC in efficiency within the MFG segment (Fusion vs Creo/Onshape).

5.4 Two-Dimensional Intersection: ADSK's "Reasonable P/E Range"

Layering the signals from both dimensions:

Weighted "Fair PE Range": 18-21x Non-GAAP Forward PE

The current 19.2x is in the lower half of the range. If the SEC discount dissipates (requires 2-3 clean quarters) + billing transition noise subsides (ongoing in FY2027) → PE could recover from 19x to 21-22x = +10-15% valuation upside potential. However, the constraint of Owner Economics (FCF-SBC Yield is only 3.4%) means that even if PE recovers, the true return might only be "moderate" rather than "generous."

5.5 Deep Dive Benchmarking: What Each Peer Company Teaches Us

After determining the "Fair PE Range," we need to understand the root causes of the differences between each peer company and ADSK in more detail – because the root causes of these differences determine whether ADSK's PE discount "should exist" or "can be remedied."

ADSK vs ADBE: Similar Growth, Different Fates

Both are large design software companies with ~12% organic growth, but ADBE Forward PE is 18x vs ADSK 19x (almost identical). On the surface, ADSK's valuation appears fair, but a deeper comparison reveals significant differences:

Lessons from ADBE: The market gives "only" 18x PE to a company with 47% OPM and a clear AI narrative – indicating that the PE ceiling for mature design software companies is approximately in the 18-22x range, regardless of profitability. For ADSK to break through this range, it needs accelerated growth or a clarified AI narrative, not just a higher OPM (already 38%).

ADSK vs NOW: The Price of Growth

NOW Forward PE 25x vs ADSK 19x – a 6x (32%) difference. Root causes:

Therefore, the 6x PE gap between NOW and ADSK almost entirely stems from differences in growth (10pp → 4x PE) and NRR (17-20pp → 1.5x PE). This is reasonable – the differences in growth and NRR are not temporary; they are determined by the business model. The ITSM market is still rapidly moving to the cloud (penetration ~15%), while the CAD/BIM market is already highly digitized (penetration >30%). ADSK cannot "become NOW" – but it doesn't need to.

ADSK vs BSY (Bentley): Purity Premium

Bentley EV/EBITDA 30.7x vs ADSK 29.7x – similar valuations, but Bentley's revenue is only $1.5B (1/5 of ADSK's). This means the market grants Bentley a significant "small and pure" premium:

Lessons from Bentley: The market is willing to pay a premium for "low SBC + high focus + clear AI narrative." If ADSK can: (1) reduce SBC from 10.9% to 7-8%, (2) clarify the APS/Neural CAD narrative, and (3) increase focus by spinning off M&E – it might achieve a similar premium. However, these are medium-to-long-term (2-3 years) possibilities, not near-term catalysts.

5.6 Benchmarking Anchor: PTC is the Most Similar Peer

Why choose PTC instead of ADBE or NOW?

PTC matches on 4 out of 5 dimensions (growth/product/customers/subscription), making it the closest peer. The revenue size difference ($2.7B vs $7.2B) implies ADSK should have a size discount (larger companies naturally grow slower), but both currently have the same PE (19x) – this suggests the market either underestimates ADSK's scale advantage or overestimates PTC.

PTC Constraint Statement: ADSK outperforms PTC in FCF Margin (33% vs 25%) and SBC burden (10.9% vs 14%). If both PEs remain identical (19x), then these advantages of ADSK are not priced in – this could be an investment opportunity, but it may also indicate that the market perceives negative factors for ADSK that PTC does not have (SEC discount + MFG competitive disadvantage).

ADSK is positioned in the "Low Growth High Profitability" quadrant—a typical Cash Cow. The market's valuation logic for Cash Cows is based on "dividend/buyback returns" rather than "growth premium." ADSK currently provides returns through share buybacks (FY2026 $1.4B, accounting for 2.7% of market capitalization). If Owner FCF Yield increases to 5%+ (requiring SBC convergence) + buybacks are maintained at 2.5% → total return >7.5% per year → attractive to Cash Cow investors.

5.7 Conclusion of This Chapter

1. Dual-dimension benchmarking confirms ADSK's "reasonable P/E range" as 18-21x. The current 19.2x is in the lower half, indicating a +10-15% P/E upside.
2. PTC is the most similar comparable. ADSK outperforms PTC in FCF and SBC, yet has the same P/E—the market has applied a combined discount to ADSK due to "SEC + Restructuring + FX" factors.
3. ADSK is a Cash Cow, not a Growth Stock—market valuation logic should focus on total return (FCF Yield + Buybacks + SBC convergence) rather than pure growth.

Chapter 6: Billing Transition Impact: How Much of FY2026's +18% is Real?

6.1 Transition Background: From Multi-Year Upfront to Annual Billing

In 2022, Autodesk announced the transition of enterprise customers' multi-year contracts from upfront billing to annual billing. This decision appears simple, but its impact on financial statements is extremely complex:

Mechanism:

• Old Model: Customer signs a 3-year contract, pays all 3 years upfront in Year 1 → Year 1 billings=$3x, Year 2-3 billings=$0
• New Model: Customer signs a 3-year contract, pays only for the current year annually → Annual billings=$1x

Impact on Financials:

• Revenue Impact Minimal: Revenue is recognized monthly according to ASC 606, regardless of when cash is received. Therefore, the timing of revenue is largely unaffected by the billing model.
• Billings Impact Significant: Old contracts billed $3x in Year 1 → New contracts bill only $1x in Year 1. Billings are "artificially" suppressed during the transition period.
• FCF Impact Significant: FCF ≈ Revenue - Expenses + Change in Deferred Revenue. A decrease in deferred revenue (due to no longer paying upfront) → FCF is "artificially" suppressed.
• Deferred Revenue Impact Significant: Non-current DR (long-term deferred revenue, representing contracts paid upfront for >12 months in the future) plummets.

This creates a three-stage distortion:

6.2 Quantifying Transition Noise: Four Dimensions

Dimension 1: Billings

Of the +30% billings growth in FY2026, management confirmed on the Q4 conference call that the new transaction model contributed ~$185M in Q4 billings and ~$107M in Q4 revenue. The annualized impact is approximately $400-500M in billings. Therefore:

FY2026 billings growth breakdown: Organic ~+18-20% + Transition catch-up ~+10-12pp = Reported +30%

FY2027 billings guidance of +9% further confirms that the catch-up effect is fading—dropping sharply from +30% to +9%.

Dimension 2: Deferred Revenue

Non-current DR plummeted from $1,377M (FY2023) to $287M (FY2026)—a 79% decrease. This is direct evidence that almost all multi-year contracts have transitioned to annual contracts. Current DR has continued to grow (+54% over five years), indicating that the core business is still expanding.

What does this tell us? Non-current DR has fallen to $287M—meaning there are not many multi-year upfront contracts remaining. Therefore, from FY2027 onwards, the catch-up effect from "old contracts expiring and being re-signed" will significantly diminish. Management also confirmed that the FY2027 transition NRR tailwind will decrease from ~3.5pp in Q1 to ~1.5pp for the full year.

Dimension 3: FCF

The sharp decline in FCF from FY2023 to FY2024 (-37%) is not due to business deterioration—rather, it is a purely mechanical effect of reduced upfront payments. FCF returned to 33.4% in FY2026 (the same level as FY2022), with FY2027 guidance at 33.8%. Therefore, 33-34% is likely ADSK's steady-state FCF Margin.

However, there's a subtle issue here: Does the FY2026 FCF ($2,409M) include "catch-up cash" (i.e., re-signing of old multi-year contracts where customers switch from owing 2-3 years to owing only 1 year, with that current year already paid)? If so, $100-200M of the FY2026 FCF of $2,409M could be a one-time catch-up—meaning the true steady-state FCF is approximately $2,200-2,300M, and FCF Margin is approximately 31-32%.

Verification: FY2027 FCF guidance is $2,700-2,800M. If FY2026 includes $200M in catch-up, then growth from $2,200M (adj) to $2,750M would be +25%, exceeding revenue growth (+12-13%). Since FCF growth should not consistently and significantly outpace revenue growth (unless margins are expanding), this suggests that FY2027 FCF may still contain some catch-up—or that ADSK's operating leverage is stronger than it appears.

Dimension 4: Revenue

Revenue is least affected by the transition (due to ASC 606 recognition), but subtle effects remain:

Management guidance for FY2027 is +12-13%—a decline of 5-6 percentage points from FY2026's +18%—which largely corresponds to the fading of the transition's contribution.

6.3 "True Organic Growth" Triple Cross-Validation

The triple validation converges on an organic growth range of 11-13%. Taking the midpoint of 12% as the growth baseline for subsequent valuation.

This compares to the 10.9% implied by a Reverse DCF: The market price is 1.1 percentage points lower than the organic growth rate—a moderate discount, not extreme.

6.4 FY2027 Outlook: The "Last Mile" of Catch-up Effect

Management provided clear guidance in their FY2027 outlook regarding the fading of transition effects:

Critical Judgment: FY2027 is a critical year for "validating true organic growth". If actual FY2027 revenue growth is <11% (below guidance)→this would prove that catch-up effects masked a deeper slowdown in growth→P/E could contract from 19x to 16-17x. If FY2027 growth is >13% (above guidance)→this would prove that organic growth momentum is stronger than market expectations→P/E could re-rate to 21-22x.

6.5 Correlation with SEC Investigation

Recalling the core findings of the SEC investigation: Management "achieved FCF and Non-GAAP OPM targets by manipulating customer billing timing (incentivizing multi-year upfront payments)". This occurred precisely against the backdrop of the billing transition—the announcement in 2022 to shift to annual payments, followed by a partial reversal in 2023 (promoting upfront payments to meet FCF targets)→SEC investigation in 2024.

This creates a reflexivity problem (where the behavior changes the observed system): We cannot be 100% certain that the FY2023 FCF of $2.02B is "real". Management may have artificially inflated FY2023 FCF by encouraging upfront payments, and then claimed the FY2024 FCF decline was "due to the transition". Both could be partially true.

Therefore, using FY2022 ($1.46B) and FY2027E ($2.75B) as "clean endpoints" is more reliable—a 5-year FCF CAGR of approximately 13.5%, and a revenue CAGR of approximately 13.2%—The two are nearly consistent (a difference of 0.3 percentage points), indicating that FCF Margin remained largely flat over these five years (33-34%). There has been no structural improvement in FCF—the FCF Margin before and after the transition is almost identical (FY2022 33.4% ≈ FY2027E 33.8%).

6.6 Chapter Conclusion: ADSK's True Picture After Stripping Away the Noise

After a four-dimensional stripping away, ADSK has transformed from "looks good but might be an illusion" to "confirmed to be decent but not exhilarating":

This "adjusted" table is crucial for subsequent valuation: Any valuation model that directly extrapolates using FY2026 reported figures (+18% growth, 33.4% FCF Margin) will systematically overvalue Autodesk. The correct baseline is: 12% organic growth + 33% FCF Margin + ~105-108% organic NRR.

6.7 Direct vs Channel Revenue Structure Drastically Changes: An Overlooked Risk

The billing transition also brought about a structural change underestimated by the market—a disruptive reorganization of sales channels:

Within two years, Autodesk transformed from being "63% channel partner-driven" to "63% direct-driven"—a 180-degree flip.

This is no coincidence. Under the new transaction model, economic transactions occur directly between Autodesk and end customers (channel partners still provide services but no longer handle cash flow). This means:

Positive Impacts:

• Autodesk directly controls customer relationships and data—previously, channel partners were the "middlemen," and Autodesk didn't know who the end customers were
• Consequently, ADSK's customer concentration plummeted from TD Synnex 39% (FY2024) to 14% (FY2026)—this is not due to distributors losing customers, but rather a shift in revenue attribution from distributors to end customers
• Direct relationships → better upsell opportunities → potential NRR improvement

Negative Risks:

• Channel partners may reduce promotion of Autodesk products due to compressed margins—if the incentive structure for 1,170 channel partners is misaligned, it could affect SMB customer acquisition
• Direct sales require greater S&M investment—the jump in FY2026 S&M expenses to $2,373M (+19% YoY) may partly stem from this
• Management announced that they will stop disclosing Direct/Indirect breakdown in FY2027—reduced transparency may conceal friction post-transition

Causal Chain: Billing Transition → Direct Economic Transactions → Channel Partner Margins Compressed → If Channel Partner Promotion Declines → SMB Customer Acquisition May Suffer → Magic Number (currently only 0.54x) May Further Deteriorate. Conversely: Direct Customer Relationships → Improved Upsell Efficiency → NRR May Increase from Organic ~105% to ~108-110%. FY2027 Q1/Q2 will provide preliminary answers as to which of these two effects will be larger.

6.8 ARPS (Average Revenue Per Subscription) Trend: The Hidden Engine of Price Increases

The breakdown of the number of subscriptions (seats) and ARPS (Average Revenue Per Subscription, average revenue per subscription—measuring how much revenue each license contributes annually) reveals an important growth structure:

Key Finding: FY2024 ARPS actually decreased (-1.4%)—because the Multi-user to Named-user (M2N) transition led to inflated subscription numbers without a corresponding increase in revenue (one multi-user license equals multiple named-user seats, but the total cost remains the same). The ARPS jump from FY2025 to FY2026 (+8%→+11%) reflects: (1) Price increases (3-5% annually), (2) Mix shift towards higher-value Collection suites, and (3) Fading of the one-time M2N effect.

Therefore, Autodesk's growth formula is: Revenue Growth ≈ Subscription Growth (+3-5%) + ARPS Growth (+7-10%) ≈ 10-15%.

However, this formula has a vulnerability: Net Adds are decelerating (785K→516K→disclosure ceased). If subscription growth slows from +5% to +2%, growth will entirely depend on ARPS improvement—which means relying entirely on price increases. In Chapter 10 (Tiered Pricing Power), we will assess whether this reliance on price increases is sustainable.

6.9 FY2027 ETR Normalization: EPS "False Acceleration"

An easily overlooked distortion in FY2026 was an anomaly in the Effective Tax Rate (ETR—the ratio of actual taxes paid to pre-tax profit):

FY2026 ETR jumped from 19.7% to 29.9%—resulting in an overpayment of $164M in taxes (the excess amount above the normal 20% ETR). This swallowed almost all of the OI growth: OI grew by $224M, taxes increased by $207M, and net profit only grew by $12M.

Therefore, FY2026 GAAP EPS ($5.23) was artificially suppressed. If ETR were normal (20%), GAAP NI would be ≈$1,282M, and EPS ≈$5.97—14% higher than actual. A significant portion of the FY2027 guidance GAAP EPS of $7.76-8.39 (+48-60%) comes from ETR normalizing from 29.9% back to ~20% (rather than business improvement).

Investors should be wary: how much of the significant EPS jump in FY2027 (+18-20% Non-GAAP, +48-60% GAAP) is "true growth" vs. "false growth due to ETR normalization"? Non-GAAP EPS is more reliable (as it already adjusts for tax rates), but the jump in GAAP EPS may create the illusion of "accelerated growth."

Chapter 7: Business Disaggregation + Geographic Breakdown

7.1 Three-Engine Growth Matrix: AECO is the Sole Growth Engine

Conclusion First: Autodesk's growth concentration is dangerously high—AECO (Architecture/Engineering/Construction/Operations) contributes 50% of revenue and over 60% of incremental revenue. If AECO decelerates, the entire company will stall.

Product Family Growth Matrix (FY2024-FY2026, $M):

Causal Chain: AECO's growth rate (17.8% CAGR) is 1.2 times the company average (14.5%) — because BIM mandates (government-mandated in 20+ countries) have created a baseline of involuntary digital demand. AutoCAD's growth rate (10.6%) is below the company average — because the 2D CAD market is highly penetrated (100M+ users), with growth primarily driven by price increases (+3-5% annually) rather than new customers. M&E (6.1%) is almost stagnant — because Blender (free and open-source) continues to erode the low-end 3D animation market.

Growth Rate Standard Deviation = 5.3pp (AECO 17.8% vs M&E 6.1% vs AutoCAD 10.6%). This is below the 10pp fragmentation threshold (SaaS M1 Q3 standard), but close. If AECO continues to accelerate and M&E continues to decelerate, FY2028 might trigger a fragmentation alert.

Under what conditions would this conclusion not hold true? If MFG (Fusion) growth accelerates to 20%+ (currently 16%), becoming a second high-growth engine — reducing reliance on AECO for concentrated growth. This would require Fusion to gain market share in the manufacturing CAD market (currently facing intense competition against PTC Creo/Onshape).

7.2 Design vs Make: Hidden Profit Margin Divergence

Autodesk also discloses revenue categorized as "Design" (design software) and "Make" (manufacturing/construction execution tools):

Make growth (+22%) > Design (+17%). This is a good signal — meaning Autodesk is expanding from "selling design tools" to "selling construction execution tools" (ACC/Payapps). However, Make's share is only 11%; even with higher growth, its impact on overall profit margins is limited. Make's profit margin is likely lower than Design's (containing more service components) — therefore, an increasing share of Make might slightly depress the blended margin.

7.3 Geographic Breakdown: EMEA Fastest Growth, APAC Faces FX Headwinds

Conclusion First: Autodesk is the most internationalized company in the SaaS cohort (56% international revenue). This brings BIM mandate-driven structural growth but also introduces FX risk that peers do not have.

Three Key Findings:

(1) EMEA is the fastest-growing region (+21%). This is not accidental. Europe is the region with the highest concentration of BIM mandates — UK (2016)/Germany (2020)/France (2017)/Spain (2024 fully mandatory)/Czech Republic (2027). Government-mandated BIM means that architectural/engineering firms must purchase BIM software such as Revit. Therefore, EMEA's growth rate exceeding Americas' is a direct result of BIM mandates.

(2) APAC is the weakest region (+11% reported / +14% CC). The 3pp FX headwind (gap between reported vs CC) indicates that APAC currencies have depreciated against the USD. More importantly, even the CC growth rate (+14%) is significantly lower than EMEA's (+21%) — because BIM mandate coverage in APAC is less extensive (primarily Singapore and Japan), and China/India/Southeast Asia still predominantly use 2D CAD.

(3) Americas' growth rate (+17%) includes billing transition catch-up. The U.S. was the first market where the billing transition was implemented. After stripping out the catch-up, Americas' organic growth rate might be ~12-13% — close to the company average.

Quantifying FX Impact: If the US Dollar Index (DXY) rises by 5%, ADSK's full-year revenue impact ≈ 56% × 5% × (after 50% natural hedging) ≈ 1.4% revenue drag (at the $100M level). This is not fatal, but it is enough to cause a 1pp miss in quarterly guidance — against a backdrop of a Forward P/E of 19x, a 1pp miss could lead to 2-3% stock price volatility.

7.4 Growth Matrix Intersection: Which Combination Is Most Important?

Impact on Investment Judgment: ADSK's growth quality hinges on the sustainability of AECO×EMEA (structural/policy-driven). If EMEA growth slows from +21% to +15% (BIM mandate penetration saturation), the company's overall growth will decrease from 12-13% to 10-11%—the 10.9% implied by a Reverse DCF falls precisely within this range. In other words, the market may have already priced in a scenario of slowing EMEA BIM mandate growth.

Chapter 8: Deep Dive into AEC Design: BIM Mandate + Revit Monopoly

8.1 BIM Market: Policy-Driven Involuntary Digitalization

Conclusion First: BIM (Building Information Modeling—using 3D digital models to replace 2D drawings for designing, constructing, and operating buildings) is the biggest technological transformation in the construction industry over the past 20 years. Autodesk's Revit is the de facto standard. Government BIM mandates are turning "optional questions" into "mandatory answers."

BIM Market Size and Growth:

• Construction BIM Software Market: $6.33B (2025) → $22.55B (2034), CAGR 20.1%
• Overall AEC Software Market: $12.04B (2026) → $17.98B (2031), CAGR 8.35%
• BIM growth (20.1%) significantly exceeds overall AEC (8.35%)—due to BIM mandates accelerating the migration from 2D→BIM

Quantifying BIM Mandate Driving Force:

20+ countries have implemented or announced BIM mandates. Divided by stage:

• Maturity Phase (2007-2016): Nordics (Finland/Norway/Denmark) + Singapore + UK + South Korea—BIM penetration in these markets already >60%, growth slowing to 8-10%
• Acceleration Phase (2017-2024): Germany/France/Spain/Dubai/Saudi Arabia/Japan—penetration 30-60%, growth 15-25%
• Explosion Phase (2025-2027): Shenzhen (2026)/Czech Republic (2027)/India (2026+)/Brazil (phased until 2028)—penetration <30%, representing the incremental source for the next 3 years

Causal Chain: Government mandates BIM→Construction companies must purchase BIM software→Revit is the de facto standard (Autodesk+Nemetschek+Bentley collectively account for >60% of the architectural BIM market)→ADSK AECO revenue growth. Because mandates are regulation-driven (not market preference), this portion of growth is non-cyclical—even in an economic downturn, governments will not withdraw BIM mandates.

The Other Side: BIM mandate-driven growth has a natural ceiling. Once penetration exceeds 80%, growth will revert to replacement/upgrade cycles (5-8%). Mature European markets (UK/Nordics) are already approaching this ceiling. The future of growth lies in emerging APAC markets (China/India/Southeast Asia)—but these markets currently contribute only 17% of revenue.

8.2 Revit's Monopoly: De Facto Standard But Not Unassailable

Revit's position in architectural BIM is similar to AutoCAD's in 2D CAD—the industry's default choice. However, their lock-in mechanisms differ:

Revit's network effect is stronger than AutoCAD's. Construction projects involve architects, structural engineers, MEP (Mechanical, Electrical, and Plumbing) engineers, and contractors—when one party uses Revit, other parties must use compatible tools to collaborate. This creates ecosystem lock-in: it's not one person deciding to use Revit, but the entire project team being "locked" into the Revit ecosystem.

IFC is the biggest structural threat to Revit's lock-in. IFC (Industry Foundation Classes—an open data exchange standard for the construction industry) is required as a delivery format by BIM mandates in 20+ countries. If IFC were good enough (all design information could be converted to IFC without loss), then clients could theoretically use any BIM tool (not limited to Revit) for authoring, and then export IFC for delivery.

However, the reality is: IFC is usable as a delivery format, but not sufficient as an authoring format. This is because IFC is an exchange format (similar to PDF), not an authoring format (similar to Word). You can export IFC for delivery to the government, but you cannot make design modifications within IFC. Therefore, IFC mandates compel Revit to support better IFC export—but do not eliminate the demand for Revit as an authoring tool.

This is why IFC mandates have actually strengthened BIM demand but not weakened Revit: Governments require IFC→Construction companies must use BIM software (cannot export IFC from 2D CAD)→Revit is the most prevalent BIM authoring tool→Revit benefits. Critics (such as Ondsel, UK Architects Association) argue that Autodesk "intentionally provides imperfect IFC support for Revit" to maintain RVT format lock-in—this criticism is valid, but it does not change Revit's market position in the short term.

8.3 AECO Growth Decomposition: Organic vs. Acquisition vs. Price Increase

AECO grew from $2,580M (FY2024) to $3,583M (FY2026)—an increase of $1,003M (+39% cumulative) over 2 years. Where did this growth come from?

Therefore, AECO's growth quality is "mixed": Organic seat growth (good) + price increases (good but with a ceiling) + acquisitions (to be validated) + transition catch-up (one-time). AECO growth in FY2027 may slow from +22% to +14-16% (excluding catch-up) — still healthy, but no longer a "+20% growth engine".

Chapter 9: Construction Tech Deep Dive: ACC vs Procore

9.1 Construction Tech Market: A $15B+ Digitalization Gap

Conclusion First: The construction industry is one of the least digitized industries globally (only behind agriculture and mining). Autodesk Construction Cloud (ACC) is competing for this $15B+ market but faces competition from Procore (Construction Management #1) and Oracle Primavera (established Project Management).

Why is Construction Tech Important? Of AECO's $3.58B revenue, most comes from the design side (Revit/AutoCAD). ACC (construction side) revenue is not disclosed separately — but it represents the potential for accelerated AECO growth. If the construction industry's digitalization expands from "design phase digitalization" to "construction + operations phase digitalization," ADSK's AECO TAM will double.

9.2 ACC vs Procore: A Clash of Two Approaches

Core Difference: Procore integrates upwards from "construction field management," while ADSK integrates downwards from "design BIM." They clash in the middle ground (document management/field collaboration).

Which Approach is Stronger? For BIM-intensive projects (large commercial buildings/infrastructure), ACC's "design → construction data continuity" is a decisive advantage — design changes made by engineers in Revit can be directly synced to ACC's construction schedule, which Procore cannot do. However, for projects with less BIM (residential/renovation/small commercial), Procore's pure construction management functions are more practical — these projects do not require complex BIM data flows.

Therefore, the market may bifurcate: Large BIM-intensive projects → ACC wins; Small to medium non-BIM projects → Procore wins. This bifurcation is good news for ADSK — because BIM mandates are pushing more and more projects into the "BIM-intensive" camp.

9.3 Procore's Vulnerability: NRR of 106% and Pricing Controversy

Procore's NRR is only 106% — which is low for SaaS (NOW~125%, DDOG~115%). GRR of 95% means an annual churn of 5% — which is high for an "industry standard" tool.

The root cause is pricing model controversy. Procore charges based on construction volume — the larger the project, the higher the fee. This means that for the same customer, if the project scale shrinks this year (economic slowdown/fewer new starts), the amount paid to Procore decreases. Volume-based pricing makes Procore's revenue highly correlated with the construction cycle — while ADSK's per-seat model is unaffected (regardless of project size, the cost of a Revit seat remains constant).

Therefore, ADSK's per-seat model is more resilient during a construction downturn — this is a structural advantage for ACC relative to Procore.

Chapter 10: MFG Deep Dive: Fusion vs PTC/Siemens/Dassault

10.1 MFG: ADSK's Weakest Yet Most Potential Engine

Conclusion First: MFG (Manufacturing) revenue of $1.38B (19% of total, +16% growth) is Autodesk's weakest core business — because in the manufacturing CAD/PLM market, ADSK faces three stronger competitors: PTC (Creo/Windchill), Siemens (NX/Teamcenter), and Dassault (CATIA/ENOVIA). However, Fusion 360 is the only cloud-native full-stack CAD/CAM/PLM product — if cloud migration replicates AEC's success in the MFG sector, MFG could become a second growth engine.

10.2 Manufacturing CAD/PLM Competitive Landscape

ADSK's true positioning in MFG is "mid-range + SMB" — not a full-market competitor. In the high-end (aerospace/automotive), CATIA and NX's precision, certifications (AS9100/IATF16949), and workflow depth far surpass Fusion. ADSK should not attempt to enter this market — it's not a gap issue, but a positioning issue.

10.3 PTC: The Most Direct Competitor

PTC is ADSK's most noteworthy competitor in the MFG sector — because their product lines overlap the most (cloud CAD: Fusion vs Onshape), and their customer types are most similar (mid-range manufacturers).

PTC's Advantages (ADSK's Disadvantages):

• GAAP OPM 35.9% vs ADSK 22% (including restructuring) — PTC has completed the profitability transformation that ADSK is currently undergoing
• PLM Depth: Windchill is an enterprise-grade PLM leader (ARR $1.53B). ADSK's Fusion Manage/Upchain is mid-market level — large manufacturers would not choose it
• SBC Discipline: PTC SBC/Rev 7.9% vs ADSK 10.9% — PTC's Owner Economics are cleaner
• CAD ARR $961M vs ADSK MFG $1,379M — similar scale, but PTC is growing faster (CC CAD +8%, PLM +10%)

ADSK's Advantages in MFG:

• Price: Fusion 360 ~$2,200/year vs PTC Creo ~$8,000-15,000/year – 3-5x lower
• Cloud-Native: Fusion has been a cloud product since its inception (unlike Creo/NX which migrated from desktop to cloud)
• Ecosystem Integration: If customers use Revit (for architectural design) and Fusion (for equipment design) simultaneously → data flow

Key Question: Can Fusion upgrade from a "cheap SMB tool" to an "mid-market enterprise standard"?

Historical Analogy: When Revit was acquired by Autodesk in 2002, it was also a "cheap BIM tool," becoming an industry standard 20 years later. However, Revit's success depended on the BIM mandate (policy-driven) + the fragmented architecture industry (no strong competitors). The MFG sector has no equivalent to a "BIM mandate," and its competitors (PTC/Siemens/Dassault) are far stronger than those in the architecture industry (ArchiCAD/Bentley). Therefore, Fusion has a lower probability of replicating Revit's path – the more likely outcome is maintaining a 20-30% share in the SMB market but failing to penetrate large enterprises.

10.4 MFG Growth Rate Breakdown of 16%

Make revenue (Manufacturing + Construction Execution) growth rate of +22% is higher than Design's +17% — indicating Fusion's momentum in new customer acquisition. However, the Magic Number is only 0.54x — indicating high customer acquisition costs and room for improvement in growth efficiency.

Chapter 11: SaaS Unit Economics

11.1 Indirect Reconstruction: Piercing the Veil of Range Disclosures

Conclusion First: Autodesk's NRR (NR3) is only disclosed in ranges (100-110%/above 110%). This lack of transparency itself is a signal – if NRR were very high (e.g., NOW ~125%), management would be incentivized to disclose it precisely. Imprecise disclosure implies NRR is not impressive enough. Indirect reconstruction results: Organic NRR ≈105-108%, potentially only ~103-105% after deducting the billing transition effect.

Reconstruction Method:

Indirect Reconstruction Model

Consistency with Reported NR3: Management stated Q2-Q4 FY2026 NR3 is ">110%" – our reconstructed 108% is slightly below 110%. The 2pp gap might stem from: (1) management using constant currency (we used reported), (2) management's NR3 definition potentially excluding certain low-priced customers, (3) our overestimation of new customer contribution.

Key Judgment: Organic NRR ~105-108% – which is on the lower side for SaaS. NOW ~125%, DDOG ~115%, Bentley 109%, Procore 106%. ADSK's NRR is similar to Procore's, but lower than Bentley's – indicating ADSK's existing customer expansion capability lags peers (relying more on price increases, less on natural usage expansion).

Under what conditions would this conclusion not hold? If ADSK's NRR definition differs from our reconstruction method (e.g., only considering customers with >$10K ACV), actual NRR might be higher (>110%). However, the fact that management chose not to disclose precisely limits optimistic interpretations.

11.2 NRR Tiered Inference: Enterprise vs. Mid-Market vs. SMB

ADSK does not disclose tiered NRR. However, we can infer from indirect evidence:

SMB tier NRR could be <100% – this is an overlooked risk. Autodesk has increased prices for 3 consecutive years + eliminated discounts (FY2024 eliminated 5% renewal discount, FY2025 eliminated multi-year discount, FY2026 multi-user repricing to 2x). For small design firms with annual budgets <$50K, an annual +5% price increase accumulating over 3 years means +16% – which could trigger migration to BricsCAD ($600/year vs AutoCAD $1,975/year).

However, the presence of Flex might mitigate SMB churn. Flex (billed daily, AutoCAD 7 tokens/day = $21/day) has an annualized cost of $5,250 – 2.7x more expensive than a subscription. However, for low-frequency users who only use it 5-10 days a month, Flex is more cost-effective than a subscription ($1,050-2,100/year vs $1,975/year). Therefore, Flex might have retained some SMB customers who would have otherwise churned – they transitioned from subscription to Flex (ARPU decreased but no churn).

11.3 Magic Number: S&M Efficiency is Low

Magic Number (Quarterly Net New ARR × 4 / Previous Quarter's S&M Expense – measures how much new annualized recurring revenue is generated for every $1 of S&M investment) is a core metric for SaaS efficiency.

A Magic Number of 0.54x is low for SaaS (DDOG~0.90x, NOW~0.75x). Reasons:

(1) S&M includes a large amount of renewal and maintenance costs. A significant portion of ADSK's $2,373M S&M is allocated to maintaining renewals for 97% of recurring revenue (channel partner commissions/customer success teams) – not entirely for new customer acquisition. If we only consider S&M spent on new customer acquisition (estimated ~40%=$950M), the new customer Magic Number ≈ $1,075M × 25% (new customer contribution) / $950M ≈ 0.28x – which is still very low.

(2) GTM restructuring is underway. Many of the 2,350 layoffs in FY2026 are from the S&M team. Management states the goal is to "shift from a renewal-focused GTM to new customer acquisition." If the restructuring is successful → the FY2027-28 Magic Number could improve to 0.6-0.7x. However, if the restructuring is merely cost-cutting → the Magic Number could further deteriorate (S&M decreases, but new customers also decrease).

(3) ADSK's growth model is driven by a "two-pronged approach of seat growth + ARPS price increases." Under this model, the Magic Number is inherently lower than for pure land-and-expand SaaS companies (e.g., DDOG's usage-based model). This is because ADSK's "expand" primarily involves price increases (without requiring additional S&M), while new seat acquisition costs are high (requiring sales/channels/training).

11.4 Rule of 40: Solid but Not Exceptional

Rule of 40(Revenue Growth % + FCF Margin % — measures the balance between growth and profitability for SaaS companies, >40% = healthy) for ADSK:

The SBC-adjusted Rule of 40 is only 36 — below the 40 benchmark. This reiterates that SBC is a core variable in ADSK's investment thesis: If SBC/Rev decreases from 10.9% to 7% (FY2029 target), Owner FCF Margin would increase by ~4pp to 28% → Rule of 40 would recover to 40.

11.5 Chapter Conclusion: The Verdict on Unit Economics

CQ3 (Pricing Power) Confidence: 55% — NRR of 105-108% indicates that price increases are executable but expansion is limited. SMB tier NRR may be <100%, and churn is already occurring. However, Flex provides a safety valve (low-frequency users convert to Flex rather than churn). Weighted pricing power is weak (drag from SMB).

CQ1 (Organic Growth) Confidence: 68% — Organic growth of 12-13% is sustainable (BIM mandate + price increases + Flex expansion). However, if NRR drops to <105%, organic growth could fall to 10-11%.

Chapter 12: Layered Assessment of AI Impact

12.1 Core Judgment: AI for ADSK is a "+3-5% Price Increase Rationale," Not a "New Revenue Curve"

Conclusion Upfront: Autodesk's AI strategy is "bundled value-add" rather than "standalone monetization" — all AI features are embedded for free into existing subscriptions, with no additional charges. This means AI will enhance retention and pricing justification in the short term (NRR↑), but it will not directly generate new revenue. Management has explicitly stated that no "discrete AI uplift" will be seen in FY2027. The real monetization window is in FY2028+, and the path depends on an API consumption model (still in early stages).

Causal Chain: Free AI bundling → enhances product stickiness → supports annual 3-5% price increases → NRR maintained at >105% (from 100-110%). If AI features are genuinely useful, customers will not forego AI assistance that saves 40% of drawing time just to save on a $2,000/year AutoCAD subscription. Therefore, AI's value to ADSK is not in the "AI revenue" line item, but in "continued pricing power" and "reduced churn" — an implicit value.

Under what conditions would this not hold true? If AI features of comparable quality are also available for free in competitors' products (BricsCAD/ZWCAD) (via open-source LLM integration), then ADSK's AI would not constitute a differentiator — bundled value-add would be zero, and the justification for price increases would disappear. If AI-native CAD solutions like Zoo.dev mature, they could bypass traditional CAD architecture from the ground up.

12.2 Six Product Lines AI Feature List: Varying Depth

Product-Level AI Maturity Matrix (10-Point Scale):

Investment Implication: AI maturity and revenue weighting are severely mismatched. Fusion (strongest AI) accounts for only 19% of revenue. AECO (50% of revenue) core tools Revit/Civil 3D have the weakest AI capabilities (2-3/10). This means ADSK's AI narrative primarily impacts the MFG business line (19% of revenue), not the growth engine AECO (50% of revenue).

Causal Reasoning: Why is Revit's AI lagging? Because the data structure of BIM models is far more complex than 2D/3D CAD – Revit includes multiple layers of semantics such as parametric families, MEP system associativity, and structural analysis interfaces. Using LLMs to generate a 2D sketch (Fusion scenario) and using LLMs to generate a complete MEP piping route (Revit scenario) are technical challenges of entirely different magnitudes. Therefore, Revit's AI lag is not a strategic error but determined by technical difficulty.

Counterpoint: If competitors (such as Snaptrude or Archistar) achieve breakthroughs in BIM AI first, ADSK's AI narrative in AECO will be debunked – "the weakest AI on the most important 50% revenue line" will transform from a technical fact into a competitive disadvantage.

12.3 AI Monetization Paths: Three-Layer Model, But First Two Layers Contribute ≈0

Management clearly described three layers of AI monetization paths during the Q4 FY2026 earnings call:

Layer One: Bundled Value-Add (Current)

• AI features are freely embedded in all subscription products
• Does not generate incremental revenue but supports the rationale for price increases (3-5% annually)
• Purpose: NRR maintenance + churn reduction
• Management's original quote: "AI monetization will play out over time... might not show discrete uplift in FY2027"

Layer Two: API Consumption-Based (Launched Q4 FY2026)

• Machine-driven API usage billing – targeting 24/7 automated workflows
• Current status: "small and early stage"
• Management's expectation: No discrete contribution in FY2027, potential start in FY2028+
• Potential clients: Large engineering firms running parametric optimization/batch rendering/automated compliance checks

Layer Three: APS Platform Ecosystem (Long-term)

• Autodesk Platform Services as the foundation for a third-party developer ecosystem
• Developers build applications → call ADSK APIs → pay-per-use
• Benchmark: Salesforce AppExchange (7,000+ apps) vs ADSK APS (undisclosed number of apps)
• Timeline: "next-decade monetization" – meaning management itself admits it's a decade-level story

Numerical Estimation:

Causal Chain: The reason AI bundled value-add (Layer One) holds the most value is that ADSK already has 8M+ subscribers – an extra $30-40/year per person amounts to an incremental $250-350M. However, this amount will not appear in any "AI revenue" line item but will be reflected as a higher ARPS (Average Revenue Per User). Therefore, using traditional SaaS AI revenue proportion to measure ADSK's AI progress is incorrect – the correct metrics are NRR trends and ARPS growth.

Counterpoint: If NRR for FY2027-2028 does not remain >105% (excluding transition effects) due to AI improvements, then the "bundled value-add" argument will not hold – AI would merely be marketing rhetoric, not a pricing anchor.

12.4 AI Impact Layered Summary: 6 Product Lines × 5-Dimension Rating

Revenue-Weighted Overall AI Score: (6.4×19% + 4.0×25% + 2.6×35% + 4.6×10% + 1.4×5% + 6.0×5%) = 3.9/10

One-Sentence Summary: Autodesk's AI is in a state of "medium depth, uneven distribution" – strong in peripheral products (Fusion/M&E), weak in core revenue (Revit/AutoCAD). The market's AI premium should be ≈0 (Fwd PE 19x already excludes AI premium), which is actually good news – if AI truly brings pricing power, the unpriced portion in the market represents upside.

Chapter 13: Neural CAD Deep Dive

13.1 Core Judgment: Neural CAD is a Gradual Threat Over 5-8 Years, Not an Acute Risk Over 2-3 Years

Executive Summary: Autodesk's Neural CAD (evolved from Project Bernini) claims to "automate 80-90% of routine design tasks." If true, the number of CAD seats enterprises require could decrease by 30-50%—this represents ADSK's own AI product cannibalizing its core subscription revenue. However, a technology maturity assessment indicates that Neural CAD will require at least 3-5 years for commercialization, with large-scale seat replacement effects likely to materialize only after 5-8 years.

Why the claim of "80-90% automation" should be heavily discounted?

1. Technical Status: Project Bernini is trained on 10 million 3D shapes and can generate 3D meshes from text/images/sketches. However, the generated mesh is "not suitable for downstream CAD development" (as Autodesk Research itself admits)—meaning the output is not an editable parametric model, but merely a "look-alike" surface geometry.
2. Commercialization Status: As of FY2026 Q4, Neural CAD has not appeared in financial disclosures—it's not in revenue breakdowns, not in management's key performance indicators, and not in guidance. This indicates that management itself does not consider it to be financially significant within the next 12 months.
3. The Chasm from "Shape Generation" to "Engineer Replacement": Professional CAD work is not about "drawing a nice-looking 3D model"—it encompasses tolerance annotation, material properties, manufacturing constraints, compliance checks, and cross-system integration. Neural CAD has currently only completed the first step (geometry generation); subsequent steps each require industry-specific knowledge (domain-specific reasoning), not just general AI capabilities.

13.2 Technology Assessment: The Evolution Path from Project Bernini to Neural CAD

Project Bernini (2024-2025, Research Phase):

• Input: Text/2D Images/Sketches/Voxels/Point Clouds
• Output: 3D Mesh (Triangular Facet Mesh)
• Training Data: 10 Million Diverse 3D Shapes
• Limitations: Output is mesh, not B-Rep (Boundary Representation—CAD's native format). Mesh to B-Rep conversion is one of the most challenging problems in the CAD domain, with no reliable automated solution currently available.
• Status: "experimental and not available for public use"

Neural CAD (2025→, Productization Attempt):

• Progress: Claims to generate "fully editable CAD geometry"—i.e., B-Rep rather than mesh.
• Architecture: Combines cutting-edge LLM + proprietary 3D models + industry knowledge reasoning.
• Deployment: In internal testing for Fusion 360 and Forma, not yet in General Availability.
• Key Unknowns: The quality of generated B-Rep (tolerance accuracy? manufacturing feasibility? compliance?) entirely lacks third-party verification.

Technology Maturity Grading (Our Assessment):

Causal Reasoning: Neural CAD's claim of "80-90% automation" is only valid at the level of "routine, repetitive design tasks"—such as standard component placement, basic layouts, and parametric variant generation. While these tasks may account for 30-50% of an engineer's daily work, the value generated (i.e., what customers are willing to pay for) might only represent 10-20%. Therefore, even if "80% of tasks" are automated, the impact on revenue might only be a "10-15% seat reduction," not 50%.

13.3 Flywheel Paradox Modeling: Three Scenarios of Neural CAD Cannibalizing Seats

The CRM flywheel paradox (Agent success → seat reduction = accelerator simultaneously acting as a brake) also exists for ADSK, but to a different degree:

Scenario A: Mild Cannibalization (50% Probability)

• Neural CAD successfully automates 30% of routine design tasks.
• Enterprises do not reduce seats; instead, engineers focus on higher-value work (design optimization/compliance/innovation).
• Seat count remains constant, but ARPS increases (because the same seat can accomplish more → engineer productivity ↑ → enterprises are willing to pay more).
• Net Revenue Impact: +3-5% (via ARPS improvement).

Scenario B: Significant Cannibalization (35% Probability)

• Neural CAD successfully automates 50% of routine design tasks.
• Large enterprises reduce 15-25% of junior engineer seats.
• Remaining seats upgrade to more expensive Premium/Collection products.
• ARPS increases by 10-15%, partially offsetting seat losses.
• Net Revenue Impact: -5% to +2% (depending on whether ARPS can compensate).

Scenario C: Severe Cannibalization (15% Probability)

• Neural CAD + external AI tool combinations enable non-professionals to complete basic designs.
• Seat demand decreases by 30-40%.
• ARPS improvement cannot fully compensate (because the seats being replaced are precisely the low-priced ones—even raising prices for high-end seats won't make up for the total volume loss).
• Net Revenue Impact: -10% to -15%.

Probability-Weighted Impact: 50%×(+4%) + 35%×(-1.5%) + 15%×(-12.5%) = +0.6%

In Summary: After probability weighting, the net impact of Neural CAD on ADSK's revenue is nearly neutral—mild cannibalization (+) and severe cannibalization (-) roughly offset each other. However, this "neutral" conclusion is highly dependent on one assumption: ADSK controls the pace of Neural CAD's rollout itself, rather than being forced to accelerate by external competitors (Zoo.dev).

13.4 Zoo.dev Threat Assessment: Real but Limited

Zoo.dev Fundamentals:

• Founded in 2021, Los Angeles.
• Led by Sequoia Capital, co-invested by GitHub co-founders.
• Core Product: Zoo Design Studio v1.1 (released January 2026).
• Key Technology: Zookeeper session proxy (text → parametric CAD geometry) + proprietary geometry engine (independent of traditional CAD kernels).
• Differentiation: "headless infrastructure"—allows CAD manipulation via code/API, rather than GUI.
• Enterprise Strategy: Accepts NX/Creo/CATIA/SolidWorks/Autodesk files → converts to its proprietary KCL format.

Threat Level Segmentation:

Why the threat from Zoo.dev is overstated?

1. CAD is not a product, it's an ecosystem: Zoo.dev can create an excellent geometry engine, but enterprises choose CAD platforms not because of "fast drawing," but because the entire supply chain (suppliers/customers/certification bodies) uses the same format. Fusion has an integrated chain with Inventor/Vault/Prodsmart; Zoo does not.
2. MFG only accounts for 19% of ADSK's revenue: Even if Zoo completely replaces Fusion's share in the SMB market (probability <10%), the impact would only be $100-200M (1-3% of total revenue).
3. Zoo's true competitor is PTC Onshape: Both are cloud-native CAD targeting mid-market manufacturing—ADSK Fusion is an incidental competitor, not the primary target.

Therefore: Zoo.dev's threat to ADSK is concentrated in the SMB segment of MFG/Fusion (19% of revenue, potentially <5% of total revenue). The short-term (3-year) financial impact is negligible. In the long term (5-10 years), if "code-driven CAD" becomes the paradigm, it could transform the entire industry. However, this constitutes "paradigm risk" rather than "competitive risk"—if it occurs, all traditional CAD vendors (ADSK/PTC/Dassault/Siemens) would be impacted.

13.5 CQ2 Verdict: Net Impact of AI – Augmentation, Not Self-Disruption, But Requires Close Monitoring

Summary of Evidence:

CQ2 Verdict: Neural CAD is an augmentative tool rather than self-disruptive – at least within the next 3-5 years. 50% of ADSK's revenue (AECO/Revit) is largely unaffected by Neural CAD (BIM model complexity serves as a natural barrier). Among the affected 19% (MFG/Fusion), the probability-weighted net effect is close to neutral.

CQ2 (AI Impact) Confidence: 55% — The net impact of AI is slightly positive to neutral. Neural CAD commercialization is further off than market narratives suggest, and Revit/AECO (50% of revenue) is protected by BIM complexity barriers. Key to monitor: If large enterprises reduce ADSK seats due to AI efficiency gains in FY2027-2028, this assessment may change.

Chapter 14: Tiered Pricing Power Assessment

14.1 Core Judgment: The "Average Value Trap" of Pricing Power – High-End Strengthening + Low-End Erosion

Conclusion First: Autodesk's pricing power is not a single number but a pair of scissors – pricing power for F500 enterprise clients is strengthening (BIM mandate + DWG lock-in + integration depth), but pricing power for SMB clients is rapidly eroding (BricsCAD alternatives + price sensitivity + improved free tools). The NRR range of 100-110% masks this divergence: high-end potentially >115%, low-end potentially <95%.

Why is layering important? Because ADSK's cumulative price increases for 2024-2025 are 13-15%. If pricing power were uniform, all customers could absorb it. If pricing power is differentiated, price increases will accelerate low-end churn – and while low-end customers have low ARPS, their volume share could be >50%. Volume churn will ultimately drag down headline subscriber growth (management has already stopped disclosing subscriber numbers in FY2026, the timing is noteworthy).

14.2 Price Increase History Reconstruction: The Aggressive Cycle of 2024-2025

Cumulative Impact: A customer renewing in 2023 will face a total price increase of approximately 13-15% by the end of 2025 (including implicit discount elimination). For example, AutoCAD: $1,975/year → ~$2,270/year.

Causal Chain: ADSK chose to implement aggressive price increases in 2024-2025 – because this coincided with the transition period for billing transformation (new transaction model), where many customers were forced to switch from multi-year contracts to annual contracts. Raising prices during this window offered two advantages: (1) Customer attention was diverted by the transition, making them less sensitive to price increases than usual; (2) The transition itself provided a reasonable excuse for "discount structure adjustment". This was a clever tactic – but it exhausted 2-3 years of pricing power.

14.3 Four-Tier Pricing Power Stage Assessment

Stage Definitions (Referencing CRM Layering Framework):

• Stage 1: No Pricing Power (Price Taker)
• Stage 2: Weak Pricing Power (Can raise prices but leads to customer churn)
• Stage 3: Moderate Pricing Power (Can raise prices 3-5%/year, customers complain but don't leave)
• Stage 4: Strong Pricing Power (Can raise prices 5-10%/year, customers have no alternative)
• Stage 5: Monopoly Pricing Power (Can raise prices 10%+/year, customers are locked in)

Layer 1: F500/Large AEC Enterprises — Stage 4 (Strong Pricing Power)

Causal Chain: Why don't F500 companies cut ADSK? Because in a large AEC project (e.g., a $500M hospital), Revit license fees might only account for 0.1-0.3% of the project cost. Even if ADSK raises prices by 10%, the impact on total project cost is 0.01-0.03 percentage points – completely negligible. However, if switching CAD systems leads to BIM model reconstruction, it could add $1-3M in costs and 3-6 months in delays. Therefore, a 10% price increase is 100 times cheaper than switching systems – this is pricing power driven by economic asymmetry.

Under what conditions would Stage 4 decline? If IFC standards mature to the point where RVT/DWG are no longer mandatory delivery formats (i.e., governments accept pure IFC submissions)→Revit's "compliance lock-in" layer is removed→Stage drops to 3. Based on buildingSMART's advancement speed, this could take 5-10 years.

Layer 2: Mid-Market (50-500-person firms) — Stage 3 (Moderate Pricing Power)

The mid-market represents the "core" of ADSK's revenue – high volume, moderate average selling price, and moderate pricing power.

Causal Chain: Why is the mid-market's pricing power only Stage 3? Because 50-500-person firms have a higher proportion of IT budget (CAD costs might account for 5-8% of personnel costs), making them more sensitive to price increases than F500s (where it's <1%). At the same time, these companies have sufficient technical capability to evaluate alternatives (unlike SMBs without IT teams), but are not large enough to negotiate enterprise discounts with ADSK (unlike F500s with EBA agreements). They are a group that is "smart but without leverage."

Layer 3: SMB/Individual Practitioners — Stage 2 (Weak Pricing Power)

Key Customer Statement: A loyal customer of over 25 years publicly complained about "a 20% license fee increase," calling it "deeply disappointing and damaging to Autodesk's reputation." This is not an isolated case – similar complaints appeared frequently in Autodesk community forums in Q1 2025 (after discounts were revoked).

Causal Chain: The root cause of weak SMB pricing power is not a poor product – AutoCAD is still the best 2D CAD. The problem is that the 2D CAD market has undergone "good-enough-ification" – BricsCAD has reached 90% of AutoCAD's functional level in 2D drafting/LISP support/DWG compatibility. When alternatives become "90% as good, 30% of the price," SMB customers (who are most sensitive to value for money) will start switching. ODA (Open Design Alliance) has 1,200+ members and monthly SDK updates, ensuring continuous DWG compatibility for BricsCAD/ZWCAD.

Layer 4: Flex Consumption Model — Stage 3.5 (Special Pricing Power)

Flex is ADSK's tactical innovation – a pay-per-day model for low-frequency users:

Causal Chain: Why are low-frequency users willing to pay a 2.3x premium? Because their usage frequency might only be 30-50 days/year – annualized Flex cost = $21 × 40 days = $840, significantly lower than the annual subscription of $2,270. Flex's "premium" only holds true when used for 250 days; actual low-frequency users save money instead. Therefore, Flex is not a "price increase tool" but a "TAM expansion tool" – bringing previously unpaid low-frequency users into the monetization scope.

Net Effect of Pricing Power Divergence: High-end strengthening (F500 Stage 4 unchanged) + Low-end erosion (SMB Stage 2 possibly dropping to 1.5) → But OPM might counter-intuitively increase – because lost SMB customers have low ARPS ($620 AutoCAD LT) and high support costs (SMBs have higher customer service/channel commission ratios), their departure actually improves the customer structure (average quality ↑). This is completely consistent with the patterns found in CRM (F500 Stage 4 / SMB Stage 2) and ADBE (Professional price increases / Consumer erosion by Canva).

14.4 Weighted Pricing Power Stage Calculation

Weighted B4 = 3.23 – higher than CRM (~2.8, with a larger SMB weighting) but lower than MSCI (~4.5, pure institutional embedding). ADSK's pricing power is not top-tier (lacking the regulatory embedding of MSCI/MCO), but it is also not fragile – Revit's 63.5% penetration in AECO + BIM mandates provide a solid baseline.

14.5 CQ3 Verdict: Pricing Power Sustainable but Growth Will Slow

Evidence Summary:

CQ3 Verdict: ADSK's pricing power is sustainable but past its peak – the aggressive price increases of 2024-2025 (13-15%) are a recent peak, and FY2027-2028 will likely see a return to the normal pace of 3-5%/year. The high-end (F500) is stable, while the low-end (SMB) is eroding. Net effect: NRR will remain at 105-110% but will not increase further.

CQ3 (Pricing Power) Confidence: 65% — Pricing power is executable, but high-end is stronger than low-end. High-end pricing power (35% of revenue) is solid, consistent with cross-verification from CRM/ADBE. Key concern: If a F500-level public defection event occurs in FY2027 (large AEC firms switching to Bentley), the pricing power narrative will be challenged.

Chapter 15: Moat Depth Assessment

15.1 Core Judgment: The moat is migrating from "format lock-in" to "model lock-in," but "platform lock-in" has not yet been established

Conclusion First: Autodesk's moat is a three-tiered structure—the bottom layer is DWG format lock-in (40 years, billions of files), the middle layer is RVT/BIM model lock-in (parametric complexity), and the top layer is APS platform lock-in (developer ecosystem). The problem is: the bottom layer is being eroded by ODA (1,200+ members), the middle layer remains solid (63.5% penetration + BIM mandate), and the top layer is almost non-existent (zero developer/application/API data).

Causal Chain: Why is a three-tiered moat structure more dangerous than "one very thick moat"? Because in a multi-layered structure, each layer has a different time window—if the erosion of the bottom layer (DWG commoditized by ODA) completes before the top layer is established (APS matures), a "moat gap" will emerge in the interim. During this gap, competitors can leverage the open DWG ecosystem + the non-existent platform lock-in to launch attacks on the middle layer (BIM).

15.2 Layer 1: DWG Format Lock-in—Half-Life Assessment

DWG is a file format created by Autodesk 40 years ago, with billions of files in existence globally. This is ADSK's oldest and most well-known moat. However, it is being systematically eroded:

ODA (Open Design Alliance) Erosion Progress:

Autodesk joining ODA is a critical signal: Historically, ADSK has consistently opposed ODA (e.g., DWG format patent lawsuits). Becoming a strategic member in 2024—signifies that ADSK management has accepted the irreversible weakening of DWG format lock-in, and the strategic focus must shift to higher layers (RVT/APS). This is a capitulation signal.

DWG Format Half-Life by Application Scenario:

Causal Reasoning: The half-life of DWG in 2D drafting scenarios is >20 years—because even if BricsCAD's functionality reaches 90%, billions of existing DWG files create enormous "read inertia" (millions of people open old DWG files daily). However, this does not mean AutoCAD's moat is safe—clients can open/edit DWG files with BricsCAD (ODA ensures compatibility); it's merely habit and training that keeps them temporarily with AutoCAD. A downgrade from "format lock-in" to "habit lock-in" has occurred—the economic value of habit lock-in is far lower than that of format lock-in (habits can be broken with a one-time training investment, while format lock-in requires data migration).

Under what conditions would the DWG half-life accelerate? If BricsCAD or ZWCAD were to launch free migration + free training programs (similar to Office → Google Workspace migration), the switching friction for SMB customers would significantly decrease. Currently, such aggressive strategies have not been observed, but economic incentives exist (BricsCAD annual fee ~$700 vs AutoCAD $2,270, a difference of $1,570/seat/year—for a 50-seat company, this saves $78,500 annually).

15.3 Layer 2: RVT/BIM Model Lock-in—Still the Core Moat

If Layer 1 (DWG) is eroding, why shouldn't ADSK's moat rating be significantly downgraded? The answer lies in Layer 2—RVT (Revit file format) and the parametric complexity of BIM models.

RVT vs DWG: Why is RVT lock-in stronger?

Key Distinction: DWG's lock-in stemmed from its "file format"—ODA cracked this lock. RVT's lock-in comes from "model complexity"—it's not a file format issue, but rather a "knowledge graph" composed of parametric families, MEP system associations (automatic clash detection for mechanical/electrical/plumbing), structural analysis interfaces, etc. Even if the RVT format were opened, competitors would need to implement a parametric engine of equivalent sophistication to truly replace it—this would require 10+ years of engineering investment.

The IFC standard is complementary, not a replacement: The IFC (Industry Foundation Classes) standard promoted by buildingSMART is often mistakenly considered a "Revit killer." In reality:

1. IFC is a data exchange format—allowing different BIM tools to read model information from each other (e.g., Revit↔ArchiCAD↔Tekla)
2. IFC is not a design format—no one directly designs buildings in IFC. Design still occurs in native tools like Revit/ArchiCAD.
3. BIM mandates require IFC interoperability—meaning governments require 'the ability to export IFC,' not 'designing with IFC.'
4. Revit already supports IFC export (though critics argue the support is not good enough)—therefore, the IFC mandate has actually strengthened Revit's position (able to export = compliant; not using Revit but still needing to use a BIM tool = market does not shrink)

The Real Threat to Layer 2: is not IFC, but rather the ability of next-generation cloud-native BIM platforms (e.g., Snaptrude/Hypar/IrisVR) to approach Revit in complexity. Currently, none can achieve this—Revit has accumulated 20 years of parametric families + MEP logic + structural analysis capabilities, and cloud-native BIM startups starting from scratch require immense engineering investment. However, if a player (potentially Google or Trimble) develops an AI-driven "Revit alternative"—leveraging large language models + architectural code understanding to automatically generate parametric BIM models—the Layer 2 barrier could be broken within 5-10 years.

15.4 Layer 3: APS Platform Lock-in—Grand Vision, Zero Data

Autodesk Platform Services (APS) Status Quo:

In addition to joining ODA in 2024, Autodesk did another thing: repositioned APS (formerly Forge) as an open platform, launching Pay-as-You-Go pricing—lowering the barrier to entry for developers. These two actions point to the same strategy: abandoning the proprietary nature of the underlying layer (DWG format) in exchange for an open ecosystem at the top layer (APS platform).

The problem is: A platform moat requires a "chicken-and-egg" two-sided market effect—developers need to see enough enterprise users to be willing to develop, and enterprise users need to see enough applications to be willing to commit. The fact that ADSK is not disclosing developer and application data, the most probable explanation is that the data is too small to be worth disclosing.

Causal Inference: Why is APS platform lock-in difficult to establish? Because the AEC industry fundamentally differs from the SaaS industry—once SaaS API integrations are established, they are difficult to dismantle (due to data flow/workflow dependencies). However, AEC workflows are still primarily "file exchange" based (export DWG/IFC → email/share) rather than "API integration." To shift AEC companies from "file exchange" to "API native" requires a transformation of the entire industry's working methods—this is not something one company can drive; it requires industry-level digital transformation. While BIM mandates have driven digitalization, governments require "delivery of BIM models," not "API collaboration."

Estimated APS Platform Lock-in Progress:

Our Assessment: ADSK platform lock-in is between Stage 2 and 3 (~15-20% migration progress)—based on 50+ APIs being open + Pay-as-You-Go model launched + but zero application/developer data. This is consistent with the assumed "~15-20%", but with limited precision (as ADSK does not disclose key data).

15.5 Overall Moat Score: CI Load-Bearing Wall Test

ADSK Moat Load-Bearing Wall Analysis (Using Financial Infrastructure CI Framework):

Weighted Moat Score: (5×15% + 8×35% + 2×10% + 6×20% + 7×20%) = 6.35/10

Moat Score Interpretation: 6.35/10 is "medium-to-strong"—stronger than pure SaaS companies (average 4-5/10, relying on switching costs), weaker than institutionally embedded companies (MSCI/MCO 8-9/10, relying on regulatory citation). ADSK's moat "appears like a monopoly (DWG 40 years + Revit 63.5%), but core defensive layers are eroding (DWG) or undeveloped (APS)"—this is a "strong exterior, weak interior" moat structure.

Key: Moat Gap Window

Moat Gap Window ≈ 2028-2033: In this 5-year window, Layer 1 (DWG) has significantly weakened, but Layer 3 (APS) is not yet established. ADSK primarily relies on Layer 2 (RVT/BIM) for defense. This period is the window of greatest opportunity for competitors (Bentley/cloud-native BIM) to launch attacks.

Under what conditions will the gap not appear? (1) If ADSK accelerates APS development, reaching Stage 4 before 2030 (requires 1,000+ applications + significant platform revenue) → the gap window shortens to 2-3 years; (2) If Layer 2 (RVT/BIM) barriers are stronger than we estimate (parametric complexity indeed requires 15-20 years to replicate) → the gap is not fatal.

15.6 CQ6 Judgment: Moat Migration Progress ~20%, Gap is a Real Risk but Not Acute

Evidence Summary:

CQ6 Judgment: The moat migration progress from DWG (format lock-in) → RVT (model lock-in) → APS (platform lock-in) is approximately 20% — Layer 1 has begun to erode, Layer 2 remains solid, and Layer 3 is almost non-existent. This is not an an immediate crisis (Layer 2's strength protects core AECO revenue), but the 2028-2033 moat gap window is a real structural risk — if ADSK fails to accelerate APS development within this window, the moat will be downgraded from "medium-to-strong" to "medium."

CQ6 (Moat Migration) Confidence: 60% — Moat migration progress is approximately 20%, with a mid-term gap risk. ODA data clearly confirms DWG commoditization, RVT strength is supported by hard data of 63.5%, and the lack of APS data is itself a signal. To watch: If ADSK discloses APS ecosystem data (developers >10K / applications >500) in FY2027-2028, the gap risk will decrease.

Chapter 16: Flywheel Validation

16.1 Core Judgment: Flywheel Exists but Net Strength ≈ 0.6 — Not a Growth Accelerator, but a Retention Tool

Conclusion First: Autodesk's flywheel is driven by "product family cross-selling" — AutoCAD users upgrade to Revit, Revit users purchase AEC Collection, and Collection users expand to ACC (Autodesk Construction Cloud). However, Chapter 9's flywheel paradox analysis shows that if Neural CAD successfully automates part of the design work, a decrease in seat demand will weaken the "volume" basis for cross-selling. The net flywheel strength (after deducting AI cannibalization) is approximately 0.6 — positive but weak, more like a "speed bump" than a "turbocharger."

16.2 Cross-Selling Path Validation: Three Real Paths

Path 1: AutoCAD→Revit→AEC Collection (AECO Mainstay)

ARPS Validation: From AutoCAD LT ($620) to AEC Collection+ACC, annual spending per user can expand 6-8 times. This is consistent with indirect NRR reconstruction (Ch7: ~111-112%) — existing customer ARPS grows 11-12% annually, partly due to this upgrade path.

Path 2: Fusion 360 Ecosystem (MFG)

Fusion 360 ($680/year) → Fusion Manage (PLM) → Fusion Operations (Manufacturing Execution) → Prodsmart (Shop Floor Data). However, the evidence for this path is weaker than for AECO — MFG growth (+16%) is lower than AECO (+22%), and Fusion faces PTC Onshape/Zoo.dev's direct competition (Ch9).

Path 3: M&E→Flow Studio (Weak Path)

M&E ($345/month) → Flow Studio (Wonder Dynamics AI VFX) → Reality Capture. This path accounts for only 4.6% of revenue, and M&E growth is the slowest (+5%). Flow Studio is an AI highlight but its scale is negligible.

16.3 Flywheel Paradox Check: AI Cannibalization's Impact on Cross-selling

Flywheel Paradox (CRM v2.0 Framework): If AI products succeed → core product seats decrease → cross-selling base shrinks → flywheel net strength needs to deduct cannibalization.

Net Flywheel Strength Calculation: Weighted 4.6/10 × Flywheel Contribution Estimate (~3-4pp from cross-selling in NRR) = Flywheel's Net Contribution to NRR approx. 2-3pp.

In short: The flywheel is real (supported by data across all three paths), but it's not ADSK's primary growth driver—price increases (3-5pp) contribute > cross-selling (2-3pp) > new customers (1-2pp). The flywheel's value lies more in retention (upgraded customers are less likely to churn) rather than growth (cross-selling itself doesn't generate substantial incremental revenue).

Chapter 17: Competitive Landscape

17.1 Core Judgment: ADSK will not be defeated by any single competitor, but may be gradually eroded by "four-front encirclement"

Conclusion First: Autodesk faces four-front competition from Bentley (horizontal infrastructure) + PTC/Zoo (manufacturing CAD) + Procore (construction management) + AI-native (foundational disruption). No single competitor can simultaneously threaten ADSK on all fronts. However, each front is eroding market share in its respective sub-market—over 5 years, this cumulative effect could lead to ADSK devolving from a "monopolist" to the "largest but not sole player." AECO (50% of revenue) is the strongest fortress; MFG (19%) and M&E (5%) are the most exposed.

17.2 Four-Front Competitive Matrix

17.3 Bentley Systems: The Parallel Universe of Horizontal Infrastructure

Bentley is not an "ADSK killer"—the overlap between the two is smaller than the market perceives.

Causal Chain: Bentley's threat is limited—because the BIM requirements for horizontal infrastructure (roads/bridges) and vertical construction (architecture/MEP) have fundamental technical differences. Road BIM uses terrain models + linear referencing, while architectural BIM uses parametric families + MEP routing—their data models are entirely different. Therefore, Bentley cannot simply "enter the architectural market." Conversely, ADSK cannot easily enter Bentley's water infrastructure/railway market (the Innovyze acquisition was an attempt, but integration has been slow).

Bentley's SBC/Rev of only 4.8% (vs ADSK 10.9%) is noteworthy—this indicates Bentley's higher earnings quality (GAAP profit is closer to true profit). If investors begin to focus on "Owner Economics" rather than "Non-GAAP" (Ch16 analysis), Bentley might achieve a valuation premium (current EV/EBITDA 30.7x vs ADSK 29.7x).

17.4 PTC: The Closest Direct Competitor

PTC has the most direct competitive relationship with ADSK's MFG business (Fusion 360):

Causal Chain: PTC's GAAP OPM of 35.9% is significantly higher than ADSK's 21.9%—because PTC has already completed the transition ADSK is currently undergoing (from perpetual licenses to subscription + restructuring). PTC's Fwd PE (18.5x) and ADSK's (19.2x) are almost identical, but PTC's GAAP profitability is notably stronger—this means that by GAAP standards, PTC is cheaper than ADSK; by Non-GAAP standards, they are similar. For investors who prioritize Owner Economics (Ch16), PTC might be a more attractive option.

Onshape vs Fusion 360:

• Onshape: Acquired by PTC in 2019 ($470M), cloud-native CAD, focused on collaborative design
• Fusion 360: Developed by ADSK, more comprehensive (CAD+CAM+CAE+PCB)
• Competitive Landscape: The mid-market (50-500 person manufacturers) is the main battlefield for both. Onshape's collaborative features (real-time multi-user editing) are superior to Fusion in some scenarios. However, Fusion's integrated CAM (machining paths) is a differentiating advantage.
• Net Assessment: Neck and neck (5/10 threat level)—both have their strengths, no overwhelming winner in the short term

17.5 Procore: Vertical Offensive in Construction Management

Causality Chain: Procore's threat is overestimated – because Procore does not do BIM design, it only competes with ADSK ACC at the "construction execution" level. ADSK's advantage, however, lies precisely in its "design-to-construction" vertical integration – BIM models designed with Revit can be seamlessly imported into ACC, a data continuity that Procore cannot replicate. Some customers are switching from Procore to ACC (for BIM-intensive projects), while reverse flow is rare.

Procore's NRR is only 106% (vs ADSK ~111%) – indicating that Procore's customer expansion capability is weaker than ADSK's. In SaaS, NRR < 110% typically means customer stickiness is insufficient to sustain high growth.

17.6 AI-native Startups: Paradigm Threat (Low Probability/High Impact)

Chapter 9 has analyzed Zoo.dev in depth. Additional AI-native threats include:

Overall Threat: The individual threat from AI-native startups is ≤3/10. However, if AI technology achieves breakthroughs (e.g., large models reliably generating parametric BIM models) – and all these startups + tech giants like Google/Apple simultaneously enter the CAD market – the cluster threat could reach 7/10. This is a 5-10 year tail risk that does not currently impact valuation but requires monitoring.

17.7 CQ7 Verdict: AECO Fortress Solid, but Four-Front Encroachment at the Edges

CQ7 Verdict: ADSK's competitive position in AECO (50% of revenue) is secure (Revit 63.5% + BIM mandate + design-to-construction vertical integration). However, MFG (19%) faces dual competition from PTC/Zoo, and M&E (5%) faces free alternatives like Blender/AI. The cumulative effect of this four-front encroachment could transform ADSK from an "undisputed leader" to the "largest but slowly declining leader" within 5 years – with annual market share erosion of ~0.5-1.0pp.

CQ7 (Competitive Landscape) Confidence: 65% — AECO's competitive position is solid, while MFG and M&E are relatively weak. Bentley's overlap with ADSK is low (approximately 15%), Procore's NRR is weak (106%), and the individual threat from AI-native startups is low.

Chapter 18: Management Assessment

18.1 Core Judgment: Management Score 5/10 – Not Excellent, but New CFO + Restructuring are Signs of Improvement

Conclusion First: ADSK management has three issues: (1) SEC investigation revealed a history of FCF manipulation → trust damaged; (2) Integration failure of large acquisition target (PlanGrid $875M) among $2B+ acquisitions → capital allocation capability questionable; (3) CEO 8 sales, 0 purchases → lack of insider conviction. However, there are also three signs of improvement: (A) SEC investigation closed with no penalties → worst outcome avoided; (B) New CFO appointed (Dec 2024) → financial discipline rebuilding; (C) 16% restructuring → if a true transformation (not just cost-cutting), structural OPM improvement is possible. Management is not a buy reason, but it's no longer a sell reason either – the governance discount is narrowing from $30-40/share (during SEC period) to $10-15/share.

18.2 SEC Investigation Review: What Was Manipulated, What Were the Consequences

Timeline:

Specific Manipulative Actions: During FY2022-FY2024, management manipulated through (1) incentivizing customers for multi-year upfront contracts → accelerating cash flow recognition; (2) controlling collection timing and accounts payable → smoothing FCF; (3) after announcing a shift to annual billing in 2022, reversing course in 2023 to promote multi-year upfront payments → to meet FCF guidance.

Causality Chain: Why did the SEC close the case without penalties? Because what was manipulated was the timing of cash flow (when money was received) rather than revenue recognition (whether it should be recognized). GAAP revenue was not impacted – what was affected were FCF and Non-GAAP margin (performance metrics chosen by management). The SEC typically imposes less severe penalties for "Non-GAAP metric manipulation" than for "GAAP violations" (which trigger SOX 404b). ADSK's situation resembles a "performance metric game" rather than "financial fraud" – questionable morally, but not legally fatal.

Impact on Investment Judgment:

• Historical FCF data needs to be discounted: FCF for FY2022-FY2024 was subject to timing manipulation and does not represent true cash-generating ability. FY2023 FCF of $2.024B (40.4% margin) might have been overstated by 15-20% (meaning true FCF margin could be ~34-35%).
• Credibility of FY2025-FY2026 data improved: New CFO appointment + implementation of remedial measures → both the motive and ability to manipulate have decreased.
• Key Risk Signal: If similar actions recur during the new CFO's tenure → systemic governance failure → valuation should incur a permanent discount.

18.3 M&A ROIC Audit: Large Targets Deviate, Small Targets Moderate

M&A Pattern Analysis:

• Large Targets ($875M PlanGrid / $1B Innovyze): Difficult integration. PlanGrid confirmed deviation (brand abandoned + talent attrition = typical acquisition failure). Innovyze shows no acceleration evidence (included in AECO, cannot be independently assessed → a negative signal in itself: if accelerated, management would proactively disclose).
• Small Targets ($240-390M): Moderate integration. Spacemaker → Forma repositioning successful; Payapps early stage but with ACC synergy logic.

Causal Chain: Why do large targets have a high probability of failure? Because large acquisitions ($800M+) typically involve independent GTM teams + product roadmaps + culture. ADSK's integration method is "absorb brand + integrate into main product line" – this is effective for small targets (Spacemaker→Forma, small team, dissectible technology), but fatal for large targets (PlanGrid had its own sales team/customer base/product vision, and talent was lost after being absorbed).

Overall M&A ROIC Assessment: Among 5 assessable acquisitions, 1 failed (PlanGrid), 2 were moderate (BC/Spacemaker), 1 is uncertain (Innovyze), and 1 is too early (Payapps). Weighted ROIC is likely <WACC – because the two largest (PlanGrid $875M + Innovyze $1B = $1.875B, accounting for over 90% of the total $2B+) lack clear evidence of success.

Core Contradiction #5 Verdict (CEO 0 Buys + $2B Acquisitions): Management is "mediocre in capital allocation and distrusts its own valuation" – CEO's zero purchases are not due to restrictions (no trading window restrictions in the past 18 months), but may reflect a neutral judgment on the $239 valuation. The ROIC of $2B acquisitions is below WACC – money spent on acquisitions is inferior to share buybacks (ADSK repurchased $5.2B during the same period, indicating a contradiction in management's judgment regarding buybacks vs. acquisitions).

18.4 Judgment on the Nature of Layoffs: Core Contradiction #3 – "Streamlining for Health" or "Panicked and Desperate"?

Evidence Framework (Verification Method of thesis_crystallization.md): If S&M (GTM restructuring) is cut while R&D remains unchanged → true restructuring. If R&D is also significantly cut → cost-cutting.

Available Evidence:

Verdict: The restructuring is closer to "streamlining for health" rather than "panicked and desperate" – because:

1. Absolute R&D expenditure is still growing (+47% over five years) – even if some R&D personnel were laid off, the increase in total investment indicates "doing more with fewer people + AI," not "cutting R&D to save money."
2. S&M efficiency improved the most (-4.1pp) – GTM restructuring (from renewal-focused to new customer acquisition) was a clear goal, validated by data.
3. Rev/Employee grew 44% – the productivity improvement from layoffs is real.
4. However: it's a "company-wide" restructuring (not just S&M) → potentially involving R&D → R&D/Rev decrease from 25.4% to 22.8% is partly contributed by headcount reduction, not purely efficiency.

Core Contradiction #3 Verdict: Layoff nature = 70% genuine transformation + 30% cost-cutting. Net effect: Structural OPM improvement is anticipated for FY2027-2028 (Non-GAAP 38%→39-40%), but R&D output quality needs monitoring (if product innovation slows down in FY2027-2028 → verdict will be revised to "cost-cutting").

18.5 CQ9 Verdict: Management 5/10 – Not Excellent but Improving

CQ9 (Management) Confidence: 50% — Management is mediocre but shows signs of improvement. Restructuring is likely a genuine transformation (70% probability), SEC investigation concluded with no penalties, and the new CFO's appointment brings expectations of improved financial discipline. Needs monitoring: If new accounting or governance issues emerge in FY2027, the judgment will be significantly downgraded.

Chapter 19: Six Years of Financial Deep Dive

19.1 Core Judgment: Revenue growth 13.2% CAGR + OPM from 14%→25% (excluding restructuring) = Quality SaaS, but tax rate and SBC significantly discount GAAP performance

Conclusion First: ADSK's FY2022-FY2026 financial narrative is "successful subscription transition + operating leverage release" – revenue 13.2% CAGR, GAAP OPM increased from 14.1% to 21.9% (excluding restructuring = 24.9%), FCF increased from $1.46B to $2.41B. However, two hidden "taxes" made actual shareholder gains significantly lower than headline figures: (1) ETR soared from 12% to 29.9% (tax law changes); (2) SBC $788M (10.9%/Rev) diluted Owner FCF Yield to only 3.4%.

19.2 Revenue Quality Assessment: 98% Recurring + AECO Engine – but Growth Divergence Intensifies

Revenue Quality Scorecard:

Causal Chain: The main risk to revenue quality is not "insufficient recurring revenue" (97% is already extremely high), but rather growth concentration – AECO contributing 60% of incremental revenue means that if the construction spending cycle declines (rising interest rates → slowing construction activity), ADSK's growth rate will drop directly from 18% to below 10%. The BIM mandate provides a floor (government projects won't stop), but private construction (accounting for ~60% of AECO) remains cyclical.

19.3 Profitability Structure: 50% of Operating Leverage Released, 50% Yet to Be Released

Operating Leverage Decomposition (FY2022→FY2026):

S&M Efficiency is the Core Engine of ADSK's Operating Leverage: S&M/Rev decreased from 37% to 33%, contributing -4.1pp—accounting for 35% of the total leverage realized (-11.6pp). This stems from: (1) a shift in channels from indirect to direct (TD Synnex from 39%→14%), reducing channel commissions; (2) restructuring of the GTM team (layoffs primarily targeting S&M); and (3) natural renewal efficiency in the subscription model (renewals do not require a new sales cycle).

Unrealized Leverage (FY2027-2030E):

• S&M: 32.9%→28-30% (continued channel directness + AI-assisted sales) = -3 to -5pp
• G&A: 9.6%→8-9% (economies of scale) = -0.5 to -1.5pp
• R&D: 22.8%→Should Not Decrease Further (product innovation requires investment, refer to Ch14 judgment) = 0pp
• Potential GAAP OPM Ceiling (excluding restructuring/SBC/ETR anomalies): 24.9% (current)→28-32% (FY2030E)

19.4 Cash Flow Quality: Strong FCF but FY2022-FY2024 Data Needs Discounting

Causal Chain: ADSK's FCF >> NI because SBC is a non-cash expense—however, SBC's dilution to shareholders is real (Ch16 analysis). Therefore, an FCF Margin of 33% does not imply "33% of revenue belongs to shareholders"—the dilution cost of SBC needs to be deducted.

19.5 Balance Sheet: Net Cash Positive + Controllable GW Risk

Key BS Indicators:

GW Impairment Risk: Of the $4,295M in GW, the two largest components are Innovyze (~$897M) and PlanGrid (included in historical GW). If Innovyze integration fails (no evidence of revenue acceleration) → $500M-$800M impairment risk. However, GW/Assets has decreased from 42% to 34% (due to total asset growth), and impairment's impact on EV is ~2-3%—not fatal but worth monitoring.

DR Deep Signal: Non-current DR plummeted from $1,377M (FY2023) to $287M—this is hard evidence of a transition from multi-year contracts to annual contracts. Impact on FCF: Reduced multi-year prepayments → decreased front-loading of FCF → FCF becomes smoother but with lower peaks. FY2023 FCF of 40.4% might have partly resulted from a "borrowing future cash" effect of multi-year prepayments; FY2026's 33.4% is closer to a true steady state.

19.6 CQ1/CQ5 Financial Judgments

CQ1 (Organic Growth): FY2026 reported growth +18%. After deducting approximately 5-6pp for billing transition catch-up → organic growth of 12-13%. This aligns with FY2027 guidance (+12-13%)—management has already factored out the transition effect in their guidance. The 12-13% organic growth can be broken down into: price increases of 3-5% + NRR expansion of 3-4% + new customers 4-5%. This breakdown is consistent with Ch7 NRR reconstruction and Ch10 pricing power analysis.

CQ1 (Organic Growth) Confidence: 72% — Organic growth of 12-13% is sustainable for 2-3 years. FY2027 guidance aligns with organic growth estimates, and the BIM mandate provides a structural floor.

CQ5 (SBC and True Profitability): SBC/Rev converged from 12.6% to 10.9%, with FY2027E <10%. The direction of convergence is correct but slow—Owner FCF Yield is only 3.4% (vs. reported FCF Yield of 4.7%). SBC represents the largest "hidden tax" in ADSK's valuation.

CQ5 (SBC and True Profitability) Confidence: 65% — The direction of SBC convergence is correct, but Owner Economics remains stringent.

Chapter 20: Three-Tier Profitability Analysis

20.1 Core Judgment: Non-GAAP Inflated by Approximately 40%—True Profitability Lies Between GAAP and Non-GAAP

Conclusion First: ADSK's profitability has three tiers: (1) GAAP EPS of $5.23 (most conservative, includes restructuring/SBC); (2) Non-GAAP EPS of $10.43 (management recommended, excludes SBC/restructuring/amortization); (3) Owner EPS of ~$6.50 (excludes restructuring/amortization but retains SBC dilution cost). The market uses Non-GAAP P/E of 19x for valuation—but if Owner P/E of 37x is used, ADSK appears expensive. The true answer lies in the middle—but leans towards the Owner Economics end.

20.2 Three-Tier Profitability Bridge (FY2026, $M)

Three-Tiered Profitability Breakdown:

20.3 SBC's True Cost: Why It Cannot Be Simply Excluded

SBC $788M (FY2026) is not "non-cash → negligible". It has two true costs:

Cost 1: Equity Dilution

• Cumulative SBC FY2022-FY2026: $555+657+703+686+788 = ~$3,389M
• Cumulative Repurchases during the same period: $1,080+1,100+795+852+1,402 = ~$5,229M
• Net Repurchases (Repurchases - SBC): ~$1,840M → 35% of repurchases were "consumed" by SBC
• However: Repurchase amount > SBC → Net share count is decreasing. Diluted shares FY2022 ~222M → FY2026 ~215M (-3%). The dilutive effect of SBC is partially offset by repurchases, but not completely.

Cost 2: Opportunity Cost

• If SBC were replaced by cash compensation → GAAP OPM would decrease by 10.9pp (from 21.9% → 11.0%)
• But it's practically impossible to fully replace it with cash (SBC is standard compensation in the tech industry)
• Reasonable Compromise: Approximately 50% of SBC is "industry standard" (talent market pricing), and 50% is "excessive incentive" (management choice)
• SBC cost to be included: $788M × 50% = ~$394M → Owner OI = $1,578 + $216 (Restructuring) + $139 (Amortization) + $16 (Acquisition) - Dilution Adjustment ≈ **~$1,950-2,050M** → Owner OPM ~27-28%

Causal Chain: Why is the "50% rule" for SBC more reasonable than full exclusion or full retention? Because if ADSK reduced SBC to 0 (all cash), it will not be able to recruit engineers (industry standard ~8-10%). However, ADSK's 10.9% is higher than PTC (7.9%) and Bentley (4.8%) – the ~3-5pp exceeding the industry standard is management's choice, not market pricing. Therefore, excluding the "industry standard" portion and retaining the "excessive" portion as an Owner cost is the fairest treatment.

20.4 ETR Normalization: FY2026's 29.9% is an Outlier

Causal Chain: FY2026 ETR of 29.9% is a one-time peak – because ADSK made tax regime election adjustments for FDII (Foreign-Derived Intangible Income) and GILTI (Global Intangible Low-Taxed Income). FY2027 guidance ETR of ~20% means NI will significantly jump after normalization:

Estimated Impact of ETR Normalization:

• FY2026 Pre-tax income: $1,603M
• At 29.9% ETR: NI = $1,124M (Actual)
• At 20% ETR: NI = $1,282M → +14% increment (ETR effect only)
• FY2027E NI estimate: Pre-tax ~$1,900M × (1-20%) = ~$1,520M → +35% YoY (includes revenue growth + ETR normalization)

This explains why FY2027 GAAP EPS guidance of $7.76-8.39 (vs FY2026 $5.23) jumps 48-60% – approximately half of which comes from ETR normalization, and half from business growth + lower restructuring expenses.

20.5 Owner FCF Yield: Investors' True Return

Three-Tiered FCF Yield Comparison:

What does an Owner FCF Yield of 3.6% mean? If ADSK's organic growth is 12-13% + FCF Yield 3.6% = a total annualized return of 15-17% (assuming valuation remains constant). This return is reasonable within SaaS – not as cheap as CRM (FCF Yield 6%+ but only 9% growth), nor as expensive as NOW (FCF Yield 2% but 22% growth). ADSK sits in the middle of the value/growth spectrum.

Comparison:

20.6 CQ8 Initial Valuation Assessment + CQ5 Final Verdict

CQ8 (Valuation): Forward Non-GAAP P/E of 19.2x is on the lower side within the SaaS cohort, but Owner P/E of ~37x is in the middle. The market uses Non-GAAP for pricing → ADSK appears cheap; using Owner for pricing → ADSK appears fair. True Positioning: Slightly Undervalued (10-15% upside potential), provided ETR normalization + SBC convergence materialize as expected.

CQ8 (Valuation) Confidence: 70% — Moderately undervalued, with approximately 10-15% upside potential. ETR normalization (29.9%→20%) provides a one-time EPS uplift, and the total return of about 16% is mid-range within the SaaS cohort. Key concern: If SBC/Rev does not decrease to <10% (FY2027) or ETR does not drop to approximately 20%, the undervaluation argument will be invalid.

CQ5 Final Verdict: SBC has converged from 12.6% to 10.9%, which is the correct direction. FY2027E <10%. However, the current Owner FCF Yield of 3.6% means ADSK's true shareholder return is significantly lower than implied by Non-GAAP. SBC is the largest divergence factor in valuation – optimists use Non-GAAP (P/E 19x = cheap), while pessimists use Owner (P/E 37x = fair).

CQ5 (SBC and True Earnings) Confidence: 65% — The direction of SBC convergence has been confirmed, but Owner Economics remain demanding.

Chapter 21: Reverse DCF Belief Inversion

21.1 Core Judgment: The Market Has Priced ADSK with "Moderate Pessimism" – Undervalued from a Standard FCF Perspective, but Overvalued from an Owner Economics Perspective

Conclusion First: The price of $235.42 (as of March 26, 2026) contains two layers of information. From a Standard FCF perspective, the market implies a 5-year revenue CAGR of 10.9%, which is lower than management's FY2027 guidance of 12.5% – this reflects moderate pessimism. However, from an Owner Economics perspective (after deducting tax-affected SBC), the implied CAGR is as high as 18.4% – this means that if SBC does not converge, the market is actually demanding growth far exceeding historical organic growth (10-12%) to justify the current price.

The divergence of these two layers of information reveals a critical fact: The core disagreement in ADSK's valuation is not about revenue growth itself (which is a consensus), but rather which framework investors choose for pricing – Standard or Owner, and the speed of SBC convergence determines which framework ultimately prevails. If SBC/Rev converges from 10.9% to 7% (Bull scenario), the Owner implied CAGR would decrease from 18.4% to ~12% – suddenly becoming achievable. If SBC remains at 11%, the "moderate pessimism" suggested by Standard FCF is merely an illusion – the true valuation demands an impossible growth rate. However, a circular dependency exists here: ~70% of the decrease in the SBC/Rev ratio comes from revenue growth (denominator effect), and only ~7% comes from a structural convergence in the absolute amount of SBC (detailed in Supporting Section S4), meaning that SBC convergence fundamentally relies on revenue growth – these two variables are not independent.

21.2 Verification of Six Implied Beliefs, One by One

The value of Reverse DCF lies not in a single number, but in deconstructing the set of beliefs behind market pricing and then testing the fragility of each belief. We identified 6 beliefs implied by the $235 price:

Belief 1: 5-Year Revenue CAGR ~10.9%

Causal Reasoning: The implied CAGR is 1.6pp below guidance, meaning the market either doesn't fully trust management (reasonable – the 2024 SEC investigation weakened guidance credibility) or has lowered its medium-term expectations after the billing transition catches up with revenue maturity. The key is that the BIM mandate covering 20+ countries provides a structural growth floor for AECO (50% of revenue) – even if all other three business lines slow down, AECO alone can support 8-9% overall growth. Therefore, an implied CAGR of 10.9% is almost impossible to miss – this is the least fragile belief.

Counterpoint: A global recession in construction spending (spillover from China's real estate crisis + high interest rates in Europe) could suppress the actual demand release from the BIM mandate. However, this would require simultaneous recessions in multiple countries – a probability below 15%.

Belief 2: Terminal FCF Margin ~35%

Causal Reasoning: ADSK's current FCF margin has already reached 33.4%, requiring only a 1.6pp expansion to hit 35%. Considering FY2026 included $216M in restructuring charges (a 3.0pp drag on OPM), after restructuring charges normalize to zero in FY2027-FY2028, GAAP OPM will naturally recover to ~25-26%, and Non-GAAP OPM could reach 40%+. An FCF margin of 35% is almost certainly achievable – this is not an optimistic assumption, but close to the current level.

Counterpoint: If ADSK must increase R&D investment to compete with Bentley/PTC (AI race), R&D/Rev could rebound from 22.8% to 25%+, compressing margins. However, the historical trend is a continuous decline in R&D/Rev (25.4%→22.8%, a 2.6pp decrease over 5 years), as economies of scale mean R&D investment does not need to grow proportionally with revenue.

Belief 3: SBC/Rev Converges to <9%

Causal Reasoning: SBC convergence is possible – management has controlled SBC growth (+15%) below revenue growth (+18%) in FY2026, meaning SBC/Rev is being passively diluted. However, the speed of convergence is the issue: After a -2% decrease in SBC in FY2025, it rebounded by +15% in FY2026, indicating that convergence is not linear – RSU vesting cycles (typically 4 years) and talent competition dictate that SBC will not compress quickly. PTC has already achieved 7.9%, demonstrating that low SBC is achievable in the same SaaS industry, but PTC experienced a slow convergence over 15+ years (SBC/Rev was once >15% in the 2010s).

Key Figures: If ADSK maintains a 12% revenue CAGR and the absolute SBC growth rate decreases to 5%/year, SBC/Rev would reach ~8.3% by FY2030 – close to, but not achieving, 7%. To reach 7%, the absolute SBC growth rate needs to be <3% or revenue growth needs to be >14%. This is the most critical "possible but uncertain" belief – the 20pp difference between Owner Economics valuation (-18%) and Standard DCF valuation (+3.2%) almost entirely depends on this variable.

Counterpoint: AI talent competition may reverse SBC convergence. AI engineer salary inflation of 20-40% in 2025-2026 (industry survey) suggests that if ADSK is to retain core engineers in its Fusion/Neural CAD teams, SBC could rebound.

Assumption 4: WACC ~10%

WACC = 10% is a standard SaaS valuation parameter (Risk-free 4.3% + ERP 5.5% + Beta adjustment). ADSK's Beta of 1.47 is high (historical 5 years), reflecting the volatility of the billing model transition. As the transition completes, Beta may revert to 1.2-1.3, and WACC could decrease to 9-9.5%—this would imply a lower implied CAGR (only 7.1% growth needed with 9% WACC), indicating deeper market pessimism. Low Vulnerability, but the direction of WACC is favorable for ADSK (downward).

Assumption 5: Terminal Growth Rate ~3.0%

A 3.0% terminal growth rate assumes the long-term growth rate for architecture and manufacturing software is the lower bound of GDP (2.5%) + inflation (2-3%). Considering that BIM mandates are still expanding globally (India/China will only start mandatory adoption in 2026+), and digital penetration still has room for improvement, 3.0% might be conservative. Extremely Low Vulnerability.

Assumption 6: AECO Maintains 50%+ Revenue Share, Growth ≥15%

AECO is ADSK's core engine—accounting for 49.7% of revenue in FY2026, with a growth rate of +22%. The market implicitly assumes AECO's continued high growth will support an overall CAGR of 10.9%+. The vulnerability of this assumption depends on Revit's position in the BIM market: Revit's BIM market share is approximately 63.5% (architects + structural engineers + MEP), and it benefits from BIM mandate lock-in—once a country mandates BIM, architectural firms typically choose the tool with the largest market share. Therefore, the baseline for AECO growth is very solid.

Counterpoint: User dissatisfaction with Revit (performance/API limitations) is an open secret. If Graphisoft ArchiCAD or Nemetschek Allplan gains institutional adoption in a mandated market (e.g., Germany), they might erode Revit's share in local markets. However, replacing an entire firm's BIM platform is a 5-10 year decision, posing no short-term threat.

21.3 Load-Bearing Wall Vulnerability Summary

Overall Assessment: Of the 8 key assumptions (pillars), 5 have low vulnerability, 2 have medium, and 1 has high (but with an impact of only -8%). This is a "hard-to-collapse" valuation structure—the most probable risk is not the collapse of a single pillar, but rather the cumulative effect of multiple pillars slightly deteriorating simultaneously (SBC not converging + MFG slowdown + renewed management issues).

Chapter 22: 5-Year Financial Review

22.1 Core Assessment: ADSK is a SaaS company where "operating leverage is materializing"—OPM expanded from -3% to 22%, but true leverage is obscured by SBC and restructuring charges

Conclusion First: ADSK's OPM has expanded from -3.0% (FY2017) to 21.9% (FY2026) over the past 10 years, a cumulative improvement of 24.9pp—this is a textbook example of SaaS operating leverage materializing. However, the GAAP OPM of 21.9% in FY2026 was suppressed by $216M in restructuring charges (3.0pp) and $788M in SBC (10.9pp)—excluding these two items, the "Core OPM" is approximately 38%, approaching the top tier of the SaaS industry (NOW 30%, CRM 33%, DDOG 25%). Operating leverage is not "about to happen"—it has already happened, merely obscured by accounting noise.

22.2 Five-Year Financial Trends: A Three-Layer Interpretation

Layer 1: Revenue—12-18% growth is real, but needs to exclude 3-5pp of billing model transition catch-up effect.

FY2026 +18% Breakdown: Organic ~12-13% + Transition Catch-up ~3-5pp + M&A ~1pp. Therefore, the slowdown in FY2027 growth to +12-13% (management guidance +12.5%) is not a deceleration—it's a normalization after the catch-up effect dissipates. This distinction is crucial: if investors interpret +18%→+12.5% as "growth decline" and sell, it would be a misunderstanding of the billing transition mechanism.

Layer 2: Profit Margin – The 16pp GAAP vs. Non-GAAP Chasm

The widening of the FY2026 Gap to 16pp (from 14pp) is not due to deteriorating SBC (which actually improved from 11.2%→10.9%), but rather caused by a one-time $216M restructuring charge. FY2027 restructuring charges are estimated to be $135-160M (approx. 1.7-2.0pp), declining to zero in FY2028→GAAP OPM will naturally jump to ~25-26%, and the Gap will shrink back to ~13pp.

Layer 3: Cash Flow – FCF Margin Experiences "V-Shaped" Recovery

Causal Chain: The FY2023 peak→FY2024 trough→FY2026 recovery perfectly tracks the billing transition timeline. In FY2023, management pushed multi-year upfront payments (subject of later SEC investigation)→FCF artificially inflated to 40.4%. In FY2024, the shift to annual billing→deferred revenue disappears→FCF plummeted to 23.3%. By FY2026, annual billing stabilizes→FCF returns to 33.4%. 33-34% is likely ADSK's true steady-state FCF margin—FY2027 guidance implies 33.8% ($2.75B/$8.14B), validating this assessment.

Important Signal: The FCF margin of 33.4% already includes 10.9% SBC (SBC is not deducted on the cash flow statement). If we calculate Owner FCF (deducting after-tax SBC) = FCF - SBC×(1-t) = $2,409M - $788M×0.80 = $1,779M → Owner FCF Margin = 24.7%. This is the true cash shareholders can realize.

22.3 Growth Quality Audit: NRR + Net Adds + ARPS Three-Dimensional Cross-Verification

NRR (Net Revenue Retention): ADSK only discloses ranges—FY2025 full year 100-110%, FY2026 Q2+ >110%. However, indirect reconstruction (Ch7) shows organic NRR of approximately 105-108%, including:

• Price Increase Contribution: 3-5pp (FY2025 7% price increase + Flex 2.7x premium)
• Cross-selling: 2-3pp (AutoCAD→Revit→Collection path)
• Churn: -2-3pp (SMB + transition friction from new transaction models)

Net Adds Deceleration: From +785K in FY2024 to +516K in FY2025 (-34%). ADSK stopped disclosing subscription numbers from FY2026 onwards, which could mean: (1) the new transaction model blurs the definition of subscriptions (reasonable explanation); (2) the numbers are no longer favorable (bearish explanation). We lean towards (1)—because the proportion of Direct revenue increased from 37% to 63%, altering the customer count methodology.

ARPS (Average Revenue Per Subscription) Trend: $688 (FY2023)→$812 (FY2026 est), +18% over 3 years. ARPS growth > subscription growth = growth model shifts from "volume-driven" to "price-driven". This aligns with the tiered pricing power analysis in Ch10: F500 customers (35% of revenue) are in Stage 4 (proactive price increases), while SMBs (20%) are in Stage 2 (passive acceptance). The question is the elasticity of Stage 2 customers—if SMB churn accelerates, NRR could fall back to <105%.

22.4 Cyclical Positioning: AEC Construction + MFG Capital Expenditure

ADSK is not a typical "cyclical stock"—but it is dually impacted by construction cycles and corporate capital expenditure cycles.

Construction Cycle (Impacts AECO 50% of Revenue): Global construction spending is projected to fluctuate at a high level in 2025-2026. The U.S. IIJA (Infrastructure Investment and Jobs Act) unlocks $1.2T+ in federal spending, coupled with Europe's green construction transition and the Middle East's NEOM/Vision 2030. However, high interest rates suppress residential construction→ADSK benefits from commercial/infrastructure (non-residential) sub-cycles rather than overall construction. The BIM mandate gives AECO growth a "semi-structural" characteristic—even if construction spending declines by 10%, mandatory adoption still provides a baseline growth of 5-8%.

Capital Expenditure Cycle (Impacts MFG 19% of Revenue): CAD/PLM investments by MFG clients follow manufacturing CapEx. Global manufacturing PMI is fluctuating around 50 in 2025-2026 (boundary between expansion/contraction). If PMI persistently stays <50, MFG growth could decline from +16% to +8-10%. However, MFG only accounts for 19% of revenue, so even a halving of its growth would only impact overall growth by ~2pp.

Cyclical Positioning Summary: ADSK is positioned in the mid-to-upper part of the cycle—construction spending remains high (bolstered by BIM mandate), but elevated interest rates curb further upside. Organic growth is projected to be 10-12% in FY2027-FY2028, potentially decelerating to 8-10% in FY2029+ (due to slower BIM mandate adoption + ARPS base effect).

22.5 Quality Score: B5 Margin Elasticity

B5: 5/5 (Highest Score)

GAAP OPM expanded from -3.0% (FY2017) to 21.9% (FY2026) over 10 years, a cumulative +24.9pp—far exceeding the "expansion >500bps" 5/5 standard. Non-GAAP OPM expanded from ~22% (FY2017 est) to 38.0%, also +16pp. ADSK has completed a full transformation from "cash-burning growth" to an "earning machine," with margin elasticity fully demonstrating the operating leverage of the SaaS business model.

A Caveat for B5: Approximately 60% of the 10-year OPM expansion stems from the SaaS transition (perpetual→subscription→annual), which is a one-time structural change rather than continuous improvement. Future OPM expansion potential depends on: (1) the pace of SBC convergence (every 1pp drop in SBC/Rev = +1pp GAAP OPM); (2) restructuring charges decreasing to zero (+3pp in FY2028); (3) economies of scale (continuous decline in S&M/Rev). Conservative estimates suggest GAAP OPM could reach 27-29% and Non-GAAP OPM could reach 40-42% by FY2030.

Chapter 23: Capital Allocation Audit

23.1 Core Judgment: ADSK is a "Passively-Biased Capital Allocator" – Buybacks Offset SBC, M&A ROIC Questionable, Zero Dividends, Zero Breakthrough Investments

Conclusion First: Over 5 years, ADSK spent $5.2B on share buybacks, largely offsetting $3.4B of SBC dilution — this is not "returning value to shareholders" but "preventing shareholder dilution", with a net effect close to zero. $2.1B in M&A was concentrated in two deals (Innovyze $1.0B + Payapps $390M + Wonder Dynamics undisclosed), with Goodwill/Assets as high as 34.4%, but Innovyze's revenue contribution remains opaque to date — this is a "satisfactory but not outstanding" capital allocation record with a B6 score of 3.5/5.

23.2 Buyback Audit: What did $5.2B buy?

5-Year Cumulative: Buybacks $5,229M / SBC $3,389M = Buyback/SBC Ratio = 1.54x — meaning that for every $1.54 in buybacks, $1.00 was used to offset SBC, and only $0.54 was true share reduction. Based on a 5-year average share price of ~$240, the actual number of shares reduced is approximately 8-10M shares (~4%), less than 1% annualized. This is not active capital return — it is a disguise for SBC costs.

Causal Reasoning: Why does ADSK not pay dividends? Two reasons: (1) SaaS industry practice — tech companies tend to favor buybacks over dividends because buybacks can offset SBC dilution and provide stock price support; (2) ADSK's net debt only approached zero in FY2024 (FY2024 Net Debt $37M), prior to which deleveraging was prioritized. Now ADSK is in a net cash position ($118M), and the conditions for initiating dividends are ripe — but management has not yet commented.

23.3 M&A Audit: Where is the ROIC for $2.1B?

Innovyze Issue: ADSK spent approximately $1.0B in FY2022 to acquire Innovyze, a water infrastructure modeling company, increasing Goodwill by $897M. However, three years later, Innovyze's revenue contribution has never been separately disclosed — it was integrated into AECO revenue. It is impossible to verify whether this $1.0B generated positive ROIC. In comparison: ADSK's ROIC is about 18.7%. If Innovyze has a dilutive effect on total ROIC (i.e., ROIC should have been >20%), then this acquisition was value-destructive.

Payapps Rationale: The $390M acquisition of construction payment SaaS aims to add a payment closed-loop to ACC (Autodesk Construction Cloud) — covering the entire process from BIM design → construction collaboration → progress payments. This rationale is strategically sound (similar to Apple Pay for the iPhone ecosystem), but Procore is also doing the same thing (Procore Pay). Key question: will Payapps' standalone NRR and ARPS be enhanced by ADSK's pricing structure? No data available at present.

Relationship between SEC Investigation and Capital Allocation: The 2024 SEC investigation revealed that management had manipulated FCF/Non-GAAP margin — which was essentially misleading investors through capital allocation data. The former CFO was reassigned, and the new CFO, Janesh Moorjani, assumed office in December 2024. The new CFO's first full fiscal year (FY2027) is a critical window to verify whether capital allocation discipline has improved.

23.4 The Economics of SBC: An "$788M annual 'hidden tax'"

SBC is not a "non-cash expense" — it is a real dilution cost for existing shareholders.

SBC Convergence Path:

• FY2026: 10.9% → FY2027E: ~10.0%(guidance implies <10%) → FY2028E: ~9.2%(Revenue growth 12% + SBC growth 5%) → FY2029E: ~8.5% → FY2030E: ~7.8%
• SBC/Rev falling to ~8% by FY2030 is the base case scenario. A reduction to 7% would require zero absolute SBC growth (unrealistic) or accelerated revenue growth.

B6 Quality Score: 3.5/5

Chapter 24: Three-Scenario Financial Projection

24.1 Key Judgement: Probability-Weighted Fair Value $243 (Standard) / $193 (Owner) — Current Price $235 is in the "neutral to slightly positive" range

24.2 Key Assumptions for Three Scenarios

24.3 Bull Case Trigger Conditions (25% Probability)

The Bull case requires ≥2 of the following 3 conditions to occur simultaneously:

1. Accelerated AI Monetization: Neural CAD/Bernini (3D generation) transitions from a narrative option to quantifiable revenue — AI-related ARR >$500M before FY2028
2. BIM Mandate Expansion: China + India + 5 other countries implement full BIM mandates between 2026-2028 — unlocking an additional $15B+ TAM
3. SBC Convergence Exceeds Expectations: New CFO leads a rebuild of SBC discipline, with SBC/Rev falling to <9% by FY2028

Historical Baseline Rate: The probability of enterprise SaaS accelerating from 12% to 14% growth is approximately 25-30% (refer to CRM 2018-2019, DDOG 2022-2023). The probability of BIM mandate expansion is relatively high (multiple countries already in planning) but the timing is uncertain. Overall assessment deems a 25% probability reasonable.

24.4 Bear Case Trigger Conditions (25% Probability)

The Bear case requires ≥2 of the following 3 conditions to occur simultaneously:

1. MFG Market Share Loss: Fusion is eroded in the mid-market by PTC Onshape or Zoo.dev, with MFG growth declining to <5%
2. Pricing Power Peak: SMB churn accelerates (NRR<100%) after 3 consecutive years of price increases, forcing a price freeze
3. Management Recidivism: A new round of accounting/governance issues, or CEO Anagnost continues to make strategic missteps (e.g., another highly-overpriced acquisition >$1B)

Historical Baseline Rate: The probability of a company that has completed its SaaS transition seeing growth decrease from 12% to 8% is approximately 20-25%. The probability of management recidivism within 18 months after an SEC investigation is extremely low (<5%), but the long-term (5-year) risk is higher — as the root cause is cultural rather than procedural. Overall assessment deems a 25% probability reasonable.

24.5 Terminal Value Sensitivity

TV as a Percentage of EV: Bull 79.5% / Base 75.4% / Bear 67.9% — Terminal Value dominates the valuation, meaning "growth/margin 5 years from now" is more important than "performance in the next 2 years". This explains why the market's reaction to FY2027 guidance was mild — short-term performance has limited impact on ADSK's long-term value.

Chapter 25: Multi-Method Valuation Framework

25.1 Summary of Six Methods

25.2 Inter-Method Dispersion Analysis

Range: $193 — $377 | Median: $273 | Inter-Method Dispersion: ($377-$193)/$273 = 67% ⚠️

A dispersion of 67% is relatively high (ideal <30%), but can be explained by:

• Systemic Differences between DCF vs. Comparables: DCF answers "what ADSK should be worth intrinsically," while comparables reflect "what the market is willing to pay for similar companies." When the overall SaaS sector's P/E expands, comparables will systematically be higher than DCF.
• 20pp Gap between Owner vs. Standard: This is not a methodological divergence—it's the result of different treatments of SBC within the same method. The $50 difference precisely equals the capitalized value of SBC.

Convergence Range: If extreme values (PEG-adjusted $377 and Owner $193) are excluded, the range for the remaining four methods is $224-$311, with a median of $258—this suggests ADSK's reasonable valuation range is $225-$310, with a midpoint of ~$260.

25.3 SOTP (Sum-of-the-Parts Valuation) – Reference Framework

Adjustments:

• Conglomerate discount 10%: -$6.0B (loss of efficiency from multi-business portfolio)
• SBC Capitalization Discount: -$6.3B ($788M × 0.80 / 10% = After-tax SBC capitalized at 10x)

Adjusted SOTP: $47.7B → $224/share

SOTP Insights: AECO ($33.6B) accounts for 56% of the total value—if the market only considers AECO's BIM mandate-driven growth and Revit's monopolistic position, ADSK would be worth $33.6B ÷ 213M = $158/share (AECO only). The remaining three segments ($24.4B) provide an additional $114/share in value. However, MFG's ($10.4B) valuation highly depends on Fusion's ability to establish a strong position in the mid-market—if Fusion's market share is captured by PTC Onshape, MFG might only be worth $5-6B (4x EV/Rev), reducing SOTP to $200/share.

25.4 Comparable Company Valuation – Reference Framework

Why is ADSK's P/E the lowest? ADSK's Forward Non-GAAP P/E of 19.2x is the lowest among these six SaaS/CAD companies – even lower than PTC (18.5x), which has similar growth (but PTC has rumors of an acquisition premium, and its GAAP OPM of 36% is significantly higher than ADSK's 22%).

Three Possible Explanations (/4 require verification):

1. SEC Discount: Although the 2024 SEC investigation has concluded, the market may still apply a 5-10% governance discount (if P/E recovers from 19x to 21x = +$25/share).
2. SBC Discount: ADSK's SBC/Rev of 10.9% is higher than PTC's 7.9% – if the market prices using Owner Economics, ADSK's "true P/E" would be approximately 37x (vs. reported 19.2x), which is not cheap.
3. FX Risk Premium: ADSK's 56% international revenue is higher than PTC's 40%, leading to greater EPS volatility in a strong USD environment.

Causal Inference: If these three discount factors gradually fade over the next two years (more than a year since the SEC investigation concluded, SBC/Rev trending downwards, USD potentially weakening), there is room for the P/E to recover from 19x to 22-25x – implying $273-$311/share (+16%~+32%). This is the quantitative basis for the summary's "moderately undervalued + 10-15% upside potential" assessment.

25.5 Forward DCF (Forward Validation) – Reference Framework

5-year projection using Base Case parameters (Rev CAGR 12%, Terminal Margin 34%, WACC 10%, g 3%):

PV (Explicit): $12.8B | PV (Terminal): $39.4B | Total EV: $52.3B → $246/share (+5%)

Forward DCF and Reverse DCF Cross-Validation: Reverse DCF indicates "$235 implies 10.9% CAGR". Forward DCF using 12% CAGR yields $246/share (+5%). The difference is only $11 (equal to a 1.1pp CAGR difference), indicating high consistency between the two methods – if ADSK can maintain 12% growth, the current price is moderately undervalued; if it falls below 11%, the current price is fair.

25.6 Valuation Summary

Consensus Direction Across Six Methods: Out of 6 methods, 4 (67%) point to upside, and 2 point to neutral/downside. However, among the 4 pointing to upside, 2 are based on comparable P/E multiples (influenced by SaaS sector valuations), and if the SaaS sector experiences de-rating (as occurred in 2022), comparable methods would be sharply adjusted downwards.

Valuation Range: $225-$310 (excluding extremes), with a midpoint of ~$260

Key Uncertainties (/4 require resolution):

• CQ3 (Pricing Power): If pricing power peaks → NRR<105% → Base revenue decreases from 12% to 9% → Fair value decreases from $246 to ~$200
• CQ9 (Management): If M&A mistakes are repeated → Goodwill impairment + valuation discount → Fair value reduced by $20-30
• SBC Convergence: If SBC/Rev ≥ 10% persists → Owner Economics remain negative → P/E multiple difficult to expand

Chapter 26: Reinforcing SaaS Unit Economics

26.1 Magic Number: 0.71 – Healthy but Not Highly Efficient

Magic Number (a metric measuring how much incremental ARR is generated for every $1 of S&M investment) = Net New ARR / Prior Year S&M

However, the above calculation is problematic: ADSK does not disclose ARR, and using revenue increment instead of ARR increment will lead to overestimation (as it includes the release of deferred revenue). A better approach is to use the cRPO increment:

Causal Reasoning: ADSK's Non-GAAP Magic Number of 0.44 is lower than DDOG's 0.92 and NOW's 0.52-0.65. The reason is that ADSK's S&M/Revenue (32.9%) is higher than NOW's (33%) but its growth rate is lower (18% vs NOW's 21%)—ADSK's sales team is focused on driving price increases for a more mature customer base, rather than rapidly expanding to new customers. This is consistent with Chapter 7 (NRR Analysis: Slowdown in Net Adds).

Counterpoint: A low Magic Number is not necessarily a bad thing—if the driver is ARPS improvement (increased revenue per customer) rather than new customer acquisition, then S&M efficiency should be measured by "incremental net revenue generated per $1 of S&M" rather than ARR increment. ADSK's ARPS from $688 to $812 (+18% over 3 years) indicates that the ROI of S&M is good in terms of price increases.

26.2 LTV/CAC: Estimated ~5.2x—Healthy (>3x threshold)

Causal Reasoning: LTV/CAC of 5.2x is above the 3x healthy threshold, but CAC Payback of ~3.1 years is higher than NOW's 1.5-2 years and DDOG's ~1.5 years. This reflects ADSK's relatively high customer acquisition cost (long sales cycles in the architecture/engineering industry + channel partner commissions), but once a customer is acquired, retention is extremely long (high CAD/BIM switching costs). ADSK's unit economics are "front-heavy (high CAC) + back-strong (high LTV)" type—suitable for long-term investors, not for those seeking quick returns.

Counterpoint: LTV calculation highly relies on churn rate estimation (ADSK does not directly disclose it). If the churn rate is underestimated (actual 8-12% instead of 5-8%), LTV/CAC could drop to 3.5x—still healthy but with a narrower margin of safety.

26.3 Rule of 40: 51.4 (Standard) / 42.7 (Owner)—Excellent

Causal Reasoning: ADSK's standard Rule of 40 score of 51.4 is only slightly below NOW's (55.5) and DDOG's (57). However, the FY2026 growth rate (+18%) includes a 3-5 percentage point catch-up from billing model transition—if calculated using an organic growth rate of ~13%, Rule of 40 = 13% + 33.4% = 46.4 (Standard) / 37.7 (Owner). The Owner-adjusted 37.7 is below the 40 threshold—this means that if measured by true shareholder return, ADSK might temporarily fall below the Rule of 40 in FY2027 (after organic growth normalizes) until SBC converges, raising the Owner FCF margin from 24.7% to 28%+.

Investment Implications: The Rule of 40 is the "health line" for the SaaS sector—SaaS companies that fall below 40 are typically penalized by the market (P/E compression to 15-20x). ADSK's standard Rule of 40 is safe (51.4), but the Owner basis might temporarily cross the line in FY2027 (37.7)—if the market starts to focus on Owner Economics, this is a risk factor for P/E compression.

Chapter 27: Key Risk Signal Register

Key Risk Signal Philosophy: KS are not "metrics we want to monitor" — they are **"binary signals that "must change our judgment once triggered"**. KS-1 (Organic Growth Rate <8%) triggered → downgrade from "Watch" to "Neutral Watch"; KS-4 (SBC >13%) triggered → Owner Economics deteriorate → P/E compression; KS-9 (New SEC investigation) triggered → Governance system failure → Valuation discount of 20%+.

Actions for Individual Triggers:

• KS-1 or KS-6: Rating downgraded by 1 notch
• KS-9: Rating immediately downgraded to "Cautionary Watch"
• KS-10 (Positive): First CEO insider buy → a 0.5-notch rating upgrade catalyst

Synergistic Triggers: KS-4 + KS-5 simultaneously triggered (SBC↑ + OPM↓) = Systemic deterioration in profitability quality → Rating downgraded by 2 notches to "Cautionary Watch"

Chapter 28: Revenue Purity Decomposition

28.1 Three-Layer Revenue Decomposition

28.2 Indirect Inference Method for Segment OPM

ADSK does not disclose segment profits. We use a triangulation inference method:

1. Aggregate Constraint: 4 segment OPMs × Revenue Weights = Overall Non-GAAP OPM 38.0%
2. AutoCAD Benchmark: AutoCAD is a 40-year-old mature product, with the lowest R&D investment and most efficient S&M (brand-driven organic traffic) → OPM is the highest among all segments. Reference: Bentley (similar mature infrastructure software) Adj OPM 28.6%, but Bentley has amortization drag from 25 acquisitions → AutoCAD's true OPM could be 35-40%.
3. AECO Inference: AECO has the fastest growth (+22%) but requires continuous R&D (Revit/Civil 3D/ACC) and moderate S&M investment → OPM could be 25-30%.
4. MFG Inference: MFG faces competition from PTC/Siemens, with the highest R&D and S&M intensity → OPM could be 15-20%.
5. Cross-Validation: 0.25×37.5% + 0.50×27.5% + 0.19×17.5% + 0.05×7.5% + 0.02×0% = 9.4%+13.8%+3.3%+0.4%+0% = 26.9% (≈GAAP OPM ex-restructuring 24.9% + departmental allocation differences in SBC) → Reasonable.

28.3 "Profit Base" Discovery

AutoCAD (25% of revenue) could contribute 40-45% of Non-GAAP profit.

Investment Implications: If AutoCAD is cannibalized by AI (KS-Risk), ADSK would not only lose 25% of its revenue, but also approximately $670M of high-margin revenue (profit per $ revenue of $0.38 vs. $0.18 for MFG). AutoCAD is a "profit base," not a "revenue base." Protecting AutoCAD's pricing power (Ch10 Pricing Power Tiering) is twice as important for profit as it is for revenue.

Chapter 29: SBC Convergence Waterfall

29.1 Decomposing SBC Change Drivers (FY2022→FY2026)

Key Insight: 70% of the decline in SBC/Rev comes from denominator growth (revenue growth faster than SBC), with only 7% from SBC structural improvement (RSU→PSU). This implies: if revenue growth slows from +18% to +12% (FY2027 normalization), the pace of SBC/Rev decline will also slow — SBC convergence is highly dependent on revenue growth, not management discipline.

NOW Comparison: NOW's SBC convergence saw 60% from denominator growth + 30% from a slowdown in absolute growth + 10% from RSU structure optimization — similar to ADSK, but NOW demonstrated better absolute growth control (SBC grew only +28% over 4 years vs. ADSK's +42%).

29.2 Convergence Projections (FY2027-FY2031)

Terminal State: SBC/Rev of approximately 8.0-8.5% (FY2031) — approaching PTC's 7.9% but unlikely to reach BSY's 4.8%. Reason: ADSK needs to compete for AI talent with Google/Microsoft, and engineer salary inflation limits the room for SBC absolute value compression.

Red Line Scenario: If AI talent competition causes SBC growth to rebound to +12% (vs. our assumed +4-8%), FY2030 SBC/Rev could stagnate at 10%+ — Owner Economics would never converge with Standard, and P/E might be locked at <20x. This is the underlying logic for KS-4 (SBC/Rev > 13%).

Chapter 30: Second Curve Validation

30.1 MFG (Fusion 360 Ecosystem) — Second Curve Validation

Second Curve Risk: MFG's 3.5/4 pass is tenuous—Growth barely exceeds the 15% threshold, and Profitability is borderline. If Fusion faces a price war from PTC Onshape/Zoo.dev in the mid-market, Growth could fall to <15% and OPM could be compressed → MFG might be downgraded from a "second curve" to a "drag". This corresponds to KS-8 (MFG growth rate < PTC).

30.2 ACC (Construction Cloud) — Second Curve Validation

Second Curve Opportunity: ACC is a more promising second curve—larger TAM ($15B vs MFG $12B), faster growth (>30%), and Procore's (pure construction SaaS) 8.0x EV/Rev suggests an independent ACC could be worth $4-5B. However, ACC's profitability is completely opaque (ADSK does not disclose it separately)—this is an information asymmetry risk.

Chapter 31: Moat Quantification: A-Score

31.1 Core Judgment: ADSK's Moat is Not One Fortress, But Three — One Eroding, One Solid, One Yet to Be Built

Conclusion First: (Ch11) gives an A-Score of 6.35/10. The task is to verify this number—to examine the strength and durability of each moat category using quantifiable data. Conclusion: An A-Score of 6.35 is reasonable, but there's significant internal differentiation: DWG file format lock-in (a 40-year legacy) is eroding at ~2-3% per year, Revit BIM's monopoly (63.5% share) is unshakeable due to mandate protection, and the APS (Autodesk Platform Services) platform moat is still in the 0-15% construction phase. Investors are betting on: whether Revit's solidity can offset DWG's erosion until APS is built.

31.2 Quantifying Five Moat Categories Individually

Moat 1: Switching Costs — Strongest Moat, Stage 4 (7.5/10)

Causal Chain: Switching costs = ADSK moat's "lower bound". Even if Revit users are dissatisfied with performance (this is public knowledge—Revit forums are full of complaints), migration means: $2-5M in direct costs + 6-18 months of productivity loss + 10-50 plugins requiring alternatives + contract lock-in penalties. Therefore, NRR maintaining at 105-108% is not due to customer satisfaction, but because leaving is more expensive. This is consistent with "Stage 4 pricing power" (Ch10): F500 clients are not "willing" to pay a premium, but "forced" to pay a premium.

Counterpoint: If ArchiCAD or Bluebeam were to launch "one-click BIM migration tools" (similar to cloud service providers' migration suites), switching costs could significantly decrease within 2-3 years. Currently, no such tools exist—but AI might make this possible (Neural CAD's self-cannibalization risk).

Moat 2: Institutional Embedding — Unique Moat, Stage 4 (8.0/10)

BIM mandates are ADSK's most unique moat—not created by ADSK, but by government regulations creating demand that specifically points to Revit.

Causal Chain: BIM mandate→architects must use BIM→Revit share 63.5%→schools teach Revit→new generation of architects use Revit→mandate strengthens Revit's position→self-reinforcing cycle. Breaking this cycle would require: (1) mandate revocation (extremely low probability, trend is strengthening not revoking), (2) another BIM tool gaining >50% educational share (currently ArchiCAD approx. 15-20%, insufficient), (3) IFC becoming a perfectly interoperable standard (technically impossible—parametric information in BIM models is lost during conversion).

Moat 3: Economies of Scale — Moderate, Stage 3 (5.0/10)

ADSK's scale advantage is reflected in: (1) $1.64B R&D (22.8% Rev)—3.6 times PTC's ($460M) and 5.9 times Bentley's ($280M); (2) Global channel network—covering 150+ countries; (3) S&M leverage—S&M/Rev from 37%→33%, -4pp over 5 years. However, the scale advantage is limited by: ADSK is not the only large CAD company—Siemens PLM (>$5B) and Dassault ($6B) are larger in the MFG sector. Therefore, scale constitutes a moat only in the AECO+AutoCAD sector (75% of revenue), but not in MFG (19%).

Moat 4: Network Effects — Weak, Stage 1-2 (3.0/10)

ADSK is not a two-sided platform—it doesn't have a "more buyers mean more sellers" flywheel. The only network effect is file format compatibility: if the entire project team (architects + structural + MEP + contractors) uses ADSK products, collaboration friction is minimized. But this is a "coordination effect" rather than a true "network effect"—it does not strengthen exponentially with an increase in user numbers.

Moat 5: Data Flywheel — High Potential but Not Yet Built, Stage 0-1 (2.0/10)

ADSK坐拥数十亿个CAD/BIM文件的设计数据——理论上这是训练AI模型的金矿(Neural CAD/Bernini就是尝试)。但目前: (1)数据没有被系统性地用于产品差异化; (2)客户对数据隐私敏感(建筑图纸=商业机密); (3)竞品也在积累数据(PTC Onshape完全云端)。数据飞轮是"期权"不是"护城河"——价值在未来,不在今天。

数据飞轮形成的三个前提条件(均未满足):

1. 数据标准化: BIM模型格式多样(RVT/IFC/DWG/NWD),需统一为AI可训练的格式→ADSK的APS正在做,但进展约15%
2. 隐私框架: 客户必须同意数据被用于AI训练。目前ADSK的隐私政策不允许未经许可使用客户设计数据。需要"匿名化聚合"或"opt-in"机制。
3. AI产品闭环: 数据→训练→AI产品→用户使用→产生更多数据→反哺训练。目前Forma(城市规划AI)可能是最接近形成闭环的产品,但用户规模太小(估计<50,000)。

可比参考: Salesforce的Data Cloud用了5年从"数据平台"进化到"可用的AI训练源",且Salesforce有数万亿条CRM记录。ADSK的BIM数据量可能是百亿级文件(40年积累),但数据密度(每个BIM文件的参数信息量)远高于CRM记录——如果ADSK能解决隐私和标准化问题,数据飞轮的质量可能超过CRM。这是5-10年的期权,不是今天的护城河。

31.3 A-Score计算

Correction: The initial score was 6.35, which was subsequently revised down to 5.90. This adjustment stems from: (1) network effects being downgraded from "medium" to "weak" (3.0 vs 4.0); (2) a more conservative assessment of the data flywheel (2.0 vs 3.5)—as both the Five Engines (Ch24) and AI assessment (Ch27) indicate that data barriers do not yet exist.

31.4 护城河迁移时间线: 从DWG到APS的"堡垒接力赛"

Core Issue: ADSK's moat is migrating from an "old fortress" (DWG file lock-in) to a "new fortress" (APS platform + BIM data). If the migration is successful, the moat could strengthen from an A-Score of 5.9 to 7-8; if it fails (DWG erosion speed > APS development speed), it could drop to 4-5.

脆弱窗口(2028-2033): This is the period when the speed of DWG erosion might exceed the speed of APS development. During this time:

• DWG file lock-in could decrease from 60% strength to 40% (ODA members 1200+, open-source DWG read/write becoming increasingly mature).
• The APS platform could migrate from 15% to 40-50% (if ADSK executes effectively).
• If the two curves "fail to cross" during 2028-2033 (fast DWG erosion + slow APS development), ADSK's A-Score could temporarily drop below 5.0 by 2030—corresponding to a P/E compression to 16-17x.

What would shorten the fragile window: (1) Rapid growth of the APS developer ecosystem (>1000 third-party applications); (2) Successful launch of Revit Cloud; (3) AI features available only on the APS platform (artificially creating migration incentives)
What would extend the fragile window: (1) Slow APS developer adoption (<200 applications); (2) Persistent delays for Revit Cloud; (3) Competitors (Bentley iTwin) leading in platformization

Investment Implications: If your investment horizon is <3 years, a moderate undervaluation at $235 + the BIM mandate floor = a reasonable investment. If the horizon is >5 years, the probability of successful APS migration must be assessed—we estimate 60-70% (based on ADSK's R&D scale advantage and Revit user base).

31.5 品质评分: B4/B7/C2/C4/C5

第32章：Playing to Win战略一致性

32.1 五层选择评分

L1: Winning Aspiration — 7/10

ADSK's aspiration is clear: "To become the design-to-construction platform for the full lifecycle in architecture and manufacturing." This aspiration has uniqueness (no other company simultaneously covers the entire AEC + MFG value chain) and defensibility (40 years of product accumulation). Reason for deduction: The aspiration is too broad—ADSK wants to win simultaneously in three major areas: AEC, MFG, and M&E, but its resources (especially R&D) might not support fighting on three fronts.

L2: 在哪里赢 — 5/10

ADSK has 4 product lines (AECO/AutoCAD/MFG/M&E) × 3 geographical markets (Americas/EMEA/APAC) = 12 strategic units. This is too dispersed. Compared to PTC (2 lines: PLM+CAD) or Procore (1 line: Construction Management), ADSK's resource allocation faces the risk of being "spread too thin." AECO (50% of revenue) should receive more resources, but M&E (5% of revenue) is also allocated R&D—is 5% of revenue worth maintaining a separate product family? This layer incurs the heaviest point deduction.

L3: How to Win — 7/10

ADSK's sources of differentiation are twofold:

1. Full Chain Coverage: End-to-end capabilities from conceptual design (AutoCAD) → detailed design (Revit) → construction (ACC) → operations & maintenance (Tandem)—competitors only cover 1-2 links in the chain. This creates cross-selling pathways (Chapter 12 Flywheel).
2. File Format Lock-in: DWG/RVT are de facto standards, and competitors must be compatible with ADSK formats.

These two sources of differentiation are sustainable (5-10 years+) but not irreplicable—Bentley and PTC are also building end-to-end capabilities in their respective domains.

L4: Core Capabilities — 7/10

ADSK's core capabilities align with L3:

• BIM Modeling Engine (Revit/Civil 3D): Highly consistent with AECO strategy
• CAD Kernel (ACIS/ShapeManager): Supports AutoCAD + Fusion
• Cloud Platform (APS): Supports the L3 "end-to-end" narrative—but APS is still in early stages (~15% migration)

Point deduction: Insufficient AI capabilities. ADSK's core capabilities in AI are weak—the Neural CAD team is significantly smaller than Google DeepMind or OpenAI, and Flow Studio (acquired from Wonder Dynamics) is a small team. If AI becomes essential for L3 differentiation, L4 will be a bottleneck.

L5: Management System — 7/10

The 16% layoff and GTM reorganization in FY2026 are signals of the management system "self-correcting"—shifting from indirect channels (TD Synnex 39%→14%) to direct sales. The appointment of a new CFO (Dec 2024) and strengthened compliance after the SEC investigation also represent improvements in the management system. However, CEO Anagnost's zero insider buying and the opaque ROIC (Chapter 19) of $2B in M&A are weaknesses in the management system.

32.2 PtW Overall Assessment

32.3 A-Score × PtW Cross-Matrix

ADSK Positioning: A-Score 5.90 + PtW 33 = Boundary Area—approaching "structural difficulties" but actually more like "A fortress with clear direction but dispersed resources". Reason: ADSK has a real fortress in AECO (Revit + mandate), but resources are dispersed across MFG/M&E → PtW is dragged down by L2. If ADSK decides to scale back MFG/M&E investments and concentrate resources on AECO+AI, PtW could increase from 33 to 38-40.

Valuation Implication: PtW 33 < ASML's 48 (semiconductor benchmark). ASML's high PtW = market grants extreme premium (Fwd PE 35x+). ADSK's moderate PtW = market grants moderate valuation (Fwd PE 19x)—this is reasonable. PE expansion to 25x+ requires PtW to increase to 40+, which necessitates L2 improvement (resource focus).

PtW Improvement Path: For ADSK to go from 33→40, L2 needs to improve from 5→8 (+3 points). Achieved by: (1) Divesting M&E (5% revenue) → 4 product lines become 3 product lines → increased focus; (2) MFG shifting from "full competition" to "Fusion+AI differentiation" → improved resource allocation; (3) Geographically focusing from 3 regions to EMEA+Americas (where BIM mandates are strongest) → reducing inefficient investment in APAC. These three items may gradually occur under the leadership of the new CFO—the 16% layoffs and GTM reorganization in FY2026 are already the first steps towards L2 improvement.

Chapter 33: Five-Engine Synergy Analysis

33.1 Engine 1: Cyclical Engine — Mid-cycle to Upper, BIM Mandate Provides a Structural Floor

Score: 6/10 (Neutral to Positive)

(Ch18.4) Positioned: ADSK is at a high point in the construction cycle + superimposed BIM mandates. Key cyclical indicators as of March 2026:

• Global Construction Spending: Remaining high, supported by IIJA + Middle East Vision 2030
• Manufacturing PMI: ~50 (expansion/contraction boundary)
• Interest Rates: High but market expects rate cuts in H2 2026 → potential acceleration in construction
• BIM Mandate: New mandates launched in Shenzhen (2026) + India (2026+) + Czech Republic (2027)

Causal Chain: Rate cuts → construction loan costs ↓ → new starts ↑ → BIM software demand ↑. However, there is a 12-18 month transmission lag. Therefore, the impact of rate cuts on revenue may not be seen in FY2027 (ending Jan 2027), but rather in FY2028.

ADSK's Non-linear Relationship with the Construction Cycle: ADSK is not a purely cyclical stock—BIM mandates create a "semi-structural" growth floor. Even if construction spending declines by 10%, mandatory adoption due to mandates can still provide a 5-8% baseline growth rate. However, ADSK is not entirely immune: If construction spending declines by >20% (2008 levels), even with mandates in place, customers might delay new license purchases or downgrade to cheaper LT versions. The FCF trough (23.3%) in FY2024 is evidence of cyclical sensitivity—although the reason was a billing transition rather than a demand decline, it proves ADSK's cash flow is highly sensitive to business changes.

Impact of Manufacturing Sub-cycle on MFG: MFG (19% of revenue) follows manufacturing CapEx. Global manufacturing PMI is oscillating around 50 in 2025-2026. The key variable is US tariff policy: If tariffs push up domestic manufacturing investment → MFG benefits (onshoring = more CAD/PLM demand); If tariffs trigger a trade war → global manufacturing contracts → MFG is negatively impacted. However, MFG accounts for only 19%, so even a halving of its growth would only impact overall growth by ~2pp.

Interest Rate Transmission Path:

Total Lag: From rate cuts to ADSK revenue impact = 12-18 months. Therefore, if rate cuts begin in H2 2026 → impact will only be seen in FY2028 (ending Jan 2028). If the current share price of $235 is pricing in "the negative impact of interest rates on AECO", it is looking in the rearview mirror—this negative factor will dissipate once the rate cut cycle begins.

33.2 Engine 2: Equity Engine — RSI 12.45 Extremely Oversold, Technicals Signal Strong Reversal

Score: 8/10 (Strong Positive)

Causal Inference: RSI at 12.45 is ADSK's lowest level in the past 5 years—the last similar reading was in October 2022 (SaaS sector crash), followed by a +45% rebound in 6 months. RSI <15 typically implies: (1) selling pressure has been fully released, (2) marginal sellers are exhausted, (3) reversal probability >70% (historical baseline). However, extreme oversold conditions do not equate to an "immediate reversal"—it can stay in the oversold zone for several weeks.

Counterpoint: Extreme oversold conditions might reflect that the market "knows something"—if a negative catalyst is imminent (e.g., FY2027 guidance cut, tariff impact), a technical reversal could be delayed.

Review of Historical RSI Extreme Events:

The median 6-month return after three historical extreme RSI events was +45%. However, the sample size is small (n=3), and each macroeconomic environment was different. The closest comparable current event is October 2022: similarly a SaaS de-rating + macro panic → +45% after 6 months. If history repeats → ADSK at ~$340 after 6 months. However, "history does not predict the future" is a necessary caveat.

Price Position Analysis: At $235, the stock is down -28.5% from its 52-week high of $329, and only up +9.5% from its 52-week low of $215. Within a risk/reward framework: downside (to $215) is approximately -$20 (-8.5%), while upside (to $329) is approximately +$94 (+40%). The risk-reward ratio is about 1:4.7—even without considering fundamentals, the pure price position offers an asymmetrical risk/reward.

33.3 Engine 3: Smart Money Engine — Persistent Insider Selling, Extremely Bullish Analysts, Diverging Signals

Rating: 4/10 (Neutral to Negative)

Insider Trading:

• 2025 Q3-Q4: 12 sells, 0 buys, net selling of 72,672 shares
• Full Year 2024: Net selling of approx. 53,000 shares (after RSU vesting)
• CEO Anagnost: 0 buys, 8 sells in 5 years
• New CFO Moorjani: No transactions (not yet vested)
• Sole buy: 1 purchase in 2025 Q1, but overwhelmed by 14 sells

Analysts:

• Ratings: 84% Buy / 16% Hold / 0% Sell
• Median Price Target: $365 (+55%↑)
• Price Target Range: $280-$400

Interpretation of Conflicting Signals: Persistent insider selling + extremely bullish analysts = classic "diverging signals". Explanation:

1. Insider selling could be for tax planning after RSU vesting (85% of SBC is RSU), not necessarily reflecting a bearish view
2. However, the CEO's zero buys is indeed a negative signal—if the CEO truly believes $235 is undervalued, why not buy?
3. Analyst bullishness might be influenced by "investment banking relationships" (ADSK is a client of frequent IPO advisors)

Conclusion: Insider selling carries more weight than analyst bullishness. If an insider buying signal emerges = strong bullish catalyst.

Deeper Interpretation of Insider Behavior: ADSK's insider trading pattern (persistent net selling) is not uncommon in the SaaS industry—executives at CRM/NOW/DDOG also regularly sell. The difference is: (1) CRM CEO Benioff sometimes makes significant purchases (high signal value); (2) NOW CEO has a shareholding commitment (>$10M). ADSK CEO Anagnost's zero buys + low shareholding (0.04% = ~$20M vs annual salary of ~$20M) implies that his economic interests are less aligned with shareholders compared to peer CEOs. This is not a "bearish" signal—rather, it's a "less committed" signal. Investment implication: If ADSK faces an acquisition offer (rumored PTC acquisition or PE buyout), the CEO might be more inclined to accept.

Institutional Holdings: Institutional ownership of ADSK is approximately 92%—highly institutionalized. The top 10 shareholders (Vanguard/BlackRock/State Street, etc.) hold ~45%, primarily index funds. This implies: (1) good liquidity (average daily trading volume ~$400M); (2) selling pressure may stem from index rebalancing rather than active judgment; (3) if an activist investor intervenes (e.g., Elliott/Starboard), the catalytic effect would be significant.

33.4 Engine 4: Signal Engine — Persistent Earnings Beat, but Growth Will Normalize

Rating: 6/10 (Neutral to Positive)

33.5 Engine 5: Prediction Market Engine — Limited Data, Only Earnings Beat Markets Available

Rating: 5/10 (Neutral)

On Polymarket, ADSK only has 3 closed markets:

1. "Will ADSK beat Q4 FY2026 earnings?" → Yes (100% settled)
2. "Will ADSK beat Q1 FY2027 earnings?" → Yes (100% settled)
3. "ADSK up or down after earnings?" → Up (100% settled)

No active forward-looking markets—this limits the informational value of the prediction market engine. There are no markets for structural questions like "Will ADSK FY2027 revenue reach $8B+" or "Will ADSK P/E return to 25x".

33.6 Overall Assessment of Five Engines Synergy

Synergistic Model: The Equity Engine (RSI extremely oversold) and Signals Engine (earnings consistently beating expectations) form a "technical + fundamental resonance"—which is typically a strong reversal signal. However, the Smart Money Engine (insider selling) provides a hedge. Net Signal: Mildly Bullish (5.9/10).

Chapter 34: PPDA Probability-Price Divergence Analysis

34.1 Three Significant Probability-Price Divergences

PPDA-1: Analyst Consensus vs. Market Pricing — 55% Divergence

Interpretation: The 55% analyst-market divergence is the largest for ADSK in the past 5 years. Historically, when ADSK's analyst-market divergence exceeded 40%, the median actual return 12 months later was approximately +25% (based on 6 events from 2019-2025). However, analyst price targets have a systematic optimistic bias (industry-wide average bias is about +15-20%), so the true expected return might be +25-35% instead of +55%.

PPDA-2: RSI Extremely Oversold vs. Fundamental Improvement — Technical/Fundamental Decoupling

Interpretation: The stock price is down -28.5% year-to-date, but FY2026 revenue is +18% and FCF has recovered to 33.4%—fundamentals have not deteriorated. The drivers of the price decline are: (1) Overall de-rating of the SaaS sector (SaaS ETF also down); (2) Tariff fears impacting international revenue expectations (ADSK 56% international); (3) "Residual discount" from the SEC investigation. If these non-fundamental factors dissipate, the price should revert to fundamentals—this is the most informative divergence.

PPDA-3: Owner Economics vs. Standard FCF — $50 Valuation Gap

Interpretation: This is not a "price-probability" divergence, but rather a methodological divergence—depending on how investors treat SBC. If the market shifts from "Standard" pricing (current) to "Owner" pricing (e.g., if SBC convergence slowing is confirmed), the stock price could fall from $235 to $193. Conversely, if SBC convergence is confirmed (FY2027 SBC/Rev < 10%), Standard and Owner will converge → supporting $240-260.

PPDA-4: Tariff Fears vs. Actual FX Impact — Emotion > Reality

Interpretation: The market's sell-off of ADSK in March 2026 (YTD -28.5%) was partly due to tariff fears—concerns that imposing tariffs on Europe/China would impact international revenue. However, ADSK's products are pure software/subscriptions, not involving the cross-border movement of physical goods—tariffs have zero direct impact on SaaS. Indirect impacts (customers' IT budgets reduced due to tariff uncertainty) might exist, but the lack of difference in FY2026 CC vs. Reported growth rates indicates that FX is not currently a drag. Therefore, the -15% decline caused by tariff fears might be an overreaction—if tariff fears subside, this portion of the decline should recover.

PPDA Overall Score: Among the 4 divergences, 3 point to "undervaluation" (PPDA-1/2/4), and 1 points to "methodological divergence" (PPDA-3). Net PPDA Direction: Leans Undervalued. However, "undervalued" does not mean "imminent reversal"—catalytic events (rate cuts/FY2027 earnings/activist involvement) might take 3-6 months to materialize.

34.2 PMSI Sentiment Index Construction

PMSI Interpretation: +1.35 = Moderately Positive (range -5 to +5). The extreme oversold condition (+4) from the Equity Engine is partially offset by downward momentum (-2) and insider selling.

PMSI and Valuation Cross-Verification: Given a valuation range of $225-310 (midpoint ~$260). PMSI +1.35 (moderately positive) is consistent with "current price of $235 being at the lower end of the range." If PMSI turns above +3 (strongly positive, requiring positive momentum + insider buying), it would signal "price reversion towards the midpoint of the range, $260-280."

Catalytic Events Ranking (by likelihood):

1. FY2027 Q1 Performance (May 2026): If Q1 beats + FY2027 guidance is not lowered → RSI reversal + analyst price target reaffirmation → PMSI may rise to +3
2. Fed Rate Cut Begins (2026 H2): Interest rates ↓ → Construction/MFG cycle ↑ → Cycle engine rises from 6 to 7-8
3. Activist Intervention: If Elliott/Starboard holds >3% → Forcing ADSK to sell M&E / focus on AECO → PtW L2 improvement → P/E expansion
4. Acquisition Rumors Materialize: PTC acquires ADSK (or reverse) → Control premium 30-50%

Most Probable 12-Month Path: Stable Q1 performance → Tariff fears subside → RSI recovery → Stock price returns to $260-280 (+10-20%). This corresponds to the "Watch" rating (+10%~+30%) forecast.

Chapter 35: AI Deep Dive: Segment-Level Impact Matrix

35.1 Core Judgment: ADSK's AI Story is "Split" – AECO is an AI Enabler, AutoCAD is AI-Vulnerable, MFG/M&E are AI-Neutral

Conclusion First: (Ch8-9) provided qualitative analysis of AI. The task of this AI deep dive is quantification. Probability-weighted AI Net Score = +1.2 (Slightly Positive) – ADSK will not be disrupted by AI, but also won't significantly benefit from AI. The biggest impact of AI on ADSK is not revenue increase or decrease, but accelerated moat migration: AI could render the DWG file format lock (a 40-year moat) irrelevant within 5-8 years, while simultaneously strengthening Revit BIM's data barrier within 3-5 years.

35.2 Layer 1: Segment-Level AI Impact Matrix

AECO (AI Enabler):

Revit + AI = generative design (automatic optimization of building layout/structure/energy consumption). ADSK's Forma (formerly Spacemaker, acquired in 2020) is already using AI for urban planning. AI enhances Revit's value – because the more complex the BIM model, the more valuable AI-assisted design becomes → A data flywheel is likely to form first in AECO: More BIM data → Better AI → Better design → More BIM adoption → More data. This is the mechanism for AECO's moat "strengthening."

Counterpoint: If AI lowers the skill barrier for BIM modeling (making it possible for non-Revit experts to produce BIM models), Revit's switching cost barrier might weaken – "Anyone can do BIM with any tool" = lower barrier to entry for competitors. But BIM is not just "modeling" – it's "parametric information management" (wall thickness, material properties, cost data, construction sequence are all linked). AI can automate modeling, not parametric management. Therefore, AI is more likely to enhance Revit (by automatically populating parameters) rather than replace it.

Quantifying AECO AI Revenue Potential:

• Forma (AI Urban Planning): Currently ~$30-50M ARR (estimated, ~$5M ARR when Spacemaker was acquired in 2022 → likely grew to $30-50M in 3 years)
• Generative Design (AI-assisted structure/layout optimization): Embedded in Revit, not separately priced → NRR improvement rather than independent revenue
• ACS (Autodesk Construction Solutions, incl. ACC+Payapps): ~$500M+ ARR (ADSK hinted in 2025 that construction cloud ARR was "approaching $1B," but it's opaque)
• Total AI Impact on AECO Revenue: If AI boosts AECO NRR from ~110% to 115% (an additional $180M in existing customer revenue per year), cumulative extra ~$2.7B over 5 years – this equates to ~3pp growth in Bull scenario revenue forecasts.

AutoCAD (AI-Vulnerable):

AutoCAD's core value is "2D/3D drafting" – which is precisely the low-level automation task AI excels at. If Neural CAD (ADSK's internal research) succeeds, it could lead to "AutoCAD + engineer" being replaced by "AI + non-professionals" → reduced seats. More dangerously, if AI drafting tools are open-source (e.g., Zoo.dev's KittyCAD) or offered free by competitors (e.g., PTC Onshape with built-in AI), AutoCAD's $2,270/seat/year pricing would lose its foundation.

Probability-weighted AI Net Score Calculation:

35.3 Layer 2: AI Implementation Depth L×S Positioning

L-Axis (Implementation Level): L1.5 (Decision Support → Controlled Automation Transition)

ADSK's AI Products:

• Forma (Urban Planning AI): L2 (Controlled Automation) – Automatically generates building layout options, human selection
• AutoCAD AI Assistant: L1 (Decision Support) – Assists with command suggestions, does not execute automatically
• Flow Studio (AI VFX): L2 – Automatically generates VFX, human review
• Neural CAD (Unreleased): L0-L1 – Conceptual stage

S-Axis (Commercial Realization): S0.5 (Narrative Option → Early Monetization Transition)

• AI-Related ARR: Not disclosed separately (= likely <$100M)
• Management AI Narrative: "AI is infused across all products" (Vague = no specific figures)
• Peer Comparison: DDOG AI revenue $100M+(S1), NOW AI commercialized $500M+(S2)

L×S Positioning: (L1.5, S0.5) — "Early Investor"

35.4 Five Invariants Test

AI Valuation Adjustment: Based on the L1.5/S0.5 positioning and a 0.5/5 invariant pass rate, AI should not contribute a significant premium to ADSK's valuation. The current $235 does not include an AI premium (Fwd PE 19x vs SaaS median 35x) – this is actually reasonable. If AI proves positive (AECO data flywheel forms), PE could expand from 19x to 22-25x (+$40-75/share). If AI proves negative (AutoCAD seat cannibalization > AECO enhancement), PE could contract from 19x to 16-17x (-$25-40/share).

35.5 AI-Adjusted Valuation

The AI probability-weighted valuation mid-point is approximately $250—consistent with the midpoint of the Standard DCF PW ($243) and comparable PE (PTC parity $273). This further confirms: $235 is moderately undervalued, but not a "bargain" – more like "fair price with a slight discount".

35.6 Technology Roadmap and Substitution Threats

ADSK's Technology Roadmap (2026-2030):

1. APS Platformization (2026-2028): API-ifying the core functionalities of AutoCAD/Revit/Fusion to allow third-party developers to build applications. The goal is to become the "AWS for AEC/MFG." Risk: APS migration is only ~15%, and the developer ecosystem is far behind Salesforce/AWS.
2. Neural CAD (2027-2029): AI-assisted/automated CAD design. Evolving from "AI suggests next action" (L1) to "AI automatically completes design tasks" (L2-L3). Risk: May cannibalize AutoCAD seats.
3. Bernini (2028+): 3D generative models (similar to DALL-E but outputting 3D models). Currently in research phase. Risk: If open-source models commercialize before Bernini → ADSK loses first-mover advantage.
4. Tandem Digital Twin (2026-2028): A bridge from BIM to operations, connecting building BIM models to IoT sensors. Strategically attractive (extending to the full building lifecycle), but the market is extremely early stage.

Ranking of Substitution Threats:

Biggest Technical Threat: Open-source AI CAD (25% probability) – if Zoo.dev's KittyCAD API + open-source AI models can achieve 80% of AutoCAD's functionality for free, AutoCAD's ($2,270/seat/year) pricing would face a fundamental challenge. However, this threat has limited impact on AECO (Revit) – BIM is an order of magnitude more complex than CAD, and open-source BIM is not realistic in the short term.

Chapter 36: Competitive Win-Loss Matrix

36.1 Battlefield Definition: ADSK Competes Simultaneously in 4 Different Markets

ADSK does not compete in "one market" – it simultaneously faces different competitors in 4 completely distinct markets. This is the root cause for PtW L2 (where to win) scoring only 5/10: each battlefield requires different product strategies, sales teams, and R&D investment.

36.2 Battlefield 1: AEC Design — Revit's "Invincible Zone" and "Erodible Zone"

Overall Assessment: ADSK holds a near-monopoly position in the AEC design sector, but this monopoly is not uniformly distributed.

"Invincible Zone" (Building Design + Structural, representing ~70% of AECO revenue):

Architectural design + structural engineering is Revit's core territory — the BIM mandate precisely covers this sector. The following causal chain explains why this sub-market is almost impossible to seize:

(1) 90%+ of architecture schools teach Revit → new graduates default to using Revit → firms recruit based on Revit skills → firms will not abandon their Revit talent pool to use ArchiCAD →Self-reinforcing talent lock-in

(2) BIM mandates require "BIM Level 2" (ISO 19650) → while not specifying Revit, the RVT format is the de facto standard for BIM model submission → approval departments also use Revit internally → if IFC format is submitted using ArchiCAD, additional format conversion and validation steps are required →Process friction prevents migration

(3) BIM collaboration for large projects (F500 clients) requires architectural + structural + MEP disciplines to all use Revit simultaneously → if one discipline switches to another tool, the model coordination costs for the entire project surge →Collaboration lock-in

Counterpoint: ArchiCAD holds >30% market share in Nordic countries (Finland/Denmark/Norway) because mandates in these countries emphasize Open BIM (IFC) and do not favor Revit. If more countries' mandates shift towards Open BIM standards → Revit's de facto standard status might decrease from 65% to 50-55% within 5-10 years. However, this is not "loss of monopoly" — it is merely "monopoly dilution."

"Erodible Zone" (Civil Infrastructure + Construction, representing ~30% of AECO revenue):

Infrastructure (Civil 3D) faces direct competition from Bentley —Bentley holds ~40% share in the roads/bridges/utilities sector, higher than ADSK's ~35%. Construction Management (ACC) faces direct competition from Procore —Procore is a pure construction SaaS, leading in document management and bid management.

36.3 Battlefield 2: Construction Technology — ACC vs Procore: "Design-side Integration" vs "Construction-side Deep Focus"

Causal Inference: ACC's competitive advantage lies in "BIM-to-Build" – if a project has already been designed using Revit, ACC can directly import the BIM model for construction management (zero friction), whereas Procore requires additional model conversion. Therefore, ACC has a structural advantage in BIM-intensive projects (large commercial/infrastructure), while Procore has a simplicity advantage in non-BIM projects (residential/small commercial).

Market Share Trend: If BIM mandates expand (increasing the proportion of BIM-intensive projects) → ACC benefits > Procore. If construction technology market growth (~20% CAGR) primarily comes from non-BIM small projects → Procore benefits > ACC. Our judgment: BIM penetration increasing from <50% → 60-70% (in 5 years) implies a rise in the proportion of BIM-intensive projects → **ACC's market positioning is superior to Procore's in the long term**.

36.4 Battlefield 3: MFG – Fusion 360's Niche Among the "Big Four"

MFG CAD/PLM is ADSK's weakest battlefield – facing PTC/Siemens/Dassault, three competitors who are larger and more focused than ADSK in the MFG domain.

Fusion's Differentiation: Fusion 360's core competitiveness is **low price + cloud-based + all-in-one functionality** (CAD+CAM+PCB+Simulation on a single platform). This holds unique appeal in the SMB and education markets (Fusion's free education version has ~10M users). However, in the Enterprise market, Fusion lacks PLM depth (Windchill/Teamcenter far surpass Fusion Manage) – therefore, MFG growth (+16%) almost entirely comes from SMB→Mid-market expansion, not from capturing Enterprise customers from PTC/Siemens.

Most Likely Outcome of MFG Competition: ADSK maintains its share in SMB/Mid (Fusion's price advantage), and does not enter Enterprise (due to insufficient PLM). MFG maintains ~19% revenue contribution, with growth gradually declining from +16% to +10-12% (SMB market saturation). **MFG will not become a growth engine – it is a stable "second pillar," contributing profit but not growth.**

Counter-argument: If ADSK acquires PTC (Engineering.com rumor), MFG would transform from a "weak competitor" to a "market leader" (Fusion+Creo+Windchill covering all segments) → PtW L2 from 5→8 → valuation re-rating. However, the acquisition probability is ~10-15%.

36.5 Battlefield 4: M&E – The Peripheral Battlefield of "Cash Cow + AI Option"

M&E (Maya/3ds Max/Flow Studio) accounts for only 5% of revenue, with the slowest growth rate (+5%). Competitors: Blender (free + open-source, rapidly growing market share), SideFX Houdini (preferred for VFX production), Cinema 4D (motion graphics).

ADSK's strategic significance in M&E is not in revenue – it's in AI data: Wonder Dynamics (acquired in 2023) → Flow Studio is ADSK's AI VFX proving ground. If AI VFX succeeds (probability ~30%), it could become a $500M+ business; if it fails, M&E would shrink to a $200-250M maintenance mode. **M&E is a "lottery ticket" – downside is limited (only 5% of revenue), upside depends on AI.**

Chapter 37: Scenario-Based Segment-Level NRR/ARPS Paths

37.1 Segment-Level Growth Drivers Breakdown

37.2 Bull Scenario (25% Probability): AI+BIM Acceleration

Bull NRR Path: AECO NRR from ~108% to 115% (AI generative design becomes a paid feature → existing customers naturally upgrade); AutoCAD NRR maintained at 105% (price increases offset AI cannibalization); MFG NRR from ~106% to 112% (Fusion PLM functionality enhancement → mid-market expansion); M&E NRR from ~102% to 108% (Flow Studio AI VFX success).

37.3 Base Scenario (50% Probability): Guidance Met

Base NRR Path: Overall NRR gradually declines from ~108% to 105% (organic, excluding FY2026 transitional catch-up). ARPS growth decreases from +12% to +6-8% (pricing power narrows). Net Adds decrease from +6-7% to +3-4% (mature market saturation).

37.4 Bear Scenario (25% Probability): Competition + AI Cannibalization

Bear NRR Path: AutoCAD NRR drops to <100% (AI-native CAD cannibalization + accelerated SMB churn) → AutoCAD revenue -12% over 5 years. MFG NRR decreases to 102% (PTC Onshape wins in mid-market). AECO NRR maintained at 105% (mandate protection → even in Bear, there's a floor).

Key Finding: In the Bear scenario, AutoCAD transforms from a "profit base" to a "contracting asset" (5Y CAGR -2%) — This is the biggest divergence point between Bear vs Base. If AutoCAD maintains +5% (+6% in Base assumption) → overall growth 9%; If AutoCAD contracts (Bear -2%) → overall growth merely 4%. The fate of AutoCAD determines the $5B revenue gap between Bull/Bear ($13.7B vs $8.7B).

Chapter 38: Moat Erosion Timeline

38.1 DWG File Format Lock-in: Half-Life Analysis

The DWG format was introduced in 1982 and has been the de facto standard for 2D CAD for 40 years. However, the Open Design Alliance (ODA, 1200+ members) is gradually eroding the exclusivity of DWG:

DWG Half-life Calculation: DWG exclusivity decreased from 95% (2000) to 60% (2026) = -35pp/26 years = -1.3pp/year. If this rate is maintained, it will drop to ~48% by 2035 – still some lock-in but no longer an "exclusive" barrier. If AI accelerates format parsing (40% probability), the rate could speed up to -2pp/year → dropping to ~38% by 2035.

Causal Inference: The direct impact of declining DWG exclusivity on ADSK is: Lower AutoCAD switching costs → SMB customers more easily switch to cheaper alternatives (Zoo.dev/FreeCAD) → AutoCAD pricing power drops from Stage 2 (passive acceptance) to Stage 1 (no pricing power) → AutoCAD ARPS growth slows from +8% to +2-3%. However, this process is slow – even if DWG exclusivity drops to 40%, Enterprise customers will not switch from AutoCAD merely because "DWG can be read" (there are other switching costs like LISP customization/plugin dependencies/training investment).

38.2 Revit BIM: Institutional Entrenchment Enhancement Curve

Unlike the erosion of DWG, Revit's moat may be strengthening:

Revit's share may peak at ~65-68% in 2028: Mandate expansion has limits – there are about 50 countries globally with construction spending >$5B, and by 2028, 30+ may have mandates → diminishing marginal returns. After 2028, IFC standardization and the Open BIM movement may begin to slowly erode Revit's share (-0.5~-1pp annually).

This means Revit's moat has a "5-year strengthening + subsequent slow dilution" non-linear path – 2026-2028 is the optimal investment window (mandates are still expanding), and 2030+ requires verifying whether APS can take over.

38.3 APS Platform: S-Curve Early Stage

APS (Autodesk Platform Services) is ADSK's "next-generation moat" – evolving from file format lock-in (DWG/RVT) to platform lock-in (API + developer ecosystem).

APS Success Probability: 60-70% (based on ADSK's R&D scale advantage + Revit user base). However, APS faces two competitors: (1) Bentley's iTwin platform (already 2000+ enterprise users, 2-3 years ahead of APS); (2) general cloud platforms (AWS/Azure) – if architectural firms choose to build on AWS themselves rather than use APS, platform lock-in would not hold.

38.4 Three-Curve Superposition: Fragile Window Probability Analysis

Vulnerable Window Probability:

• 2028-2033: DWG exclusivity drops to <50% + APS has not yet reached 30%+ = Net Moat May Temporarily Decline
• Window Probability: If APS accelerates (60% probability) → Short window (2 years); If APS stagnates (40% probability) → Long window (5+ years)
• Probability-Weighted Vulnerable Window Length: 60%×2 years + 40%×5 years = 3.2 years
• Valuation Impact During Vulnerable Window: A-Score may temporarily drop to 5.0 → P/E compresses from 19x to 16-17x → $200-215/share

Chapter 39: SBC Sensitivity Matrix

39.1 SBC/Rev Change → Owner FCF → Valuation Impact Chain

Key Finding: Each 1pp drop in SBC/Rev → Owner FCF +$58M → Valuation +$7/share (approx. +3%). From current 10.9% → 7% (Bull target) = +$27/share. From 10.9% → 12% (Bear scenario) = -$7/share.

Standard vs. Owner Convergence Conditions: When SBC/Rev drops to ~5%, the gap between Standard P/E (~19x) and Owner P/E (~22x) narrows to 3x (from current 18x gap) – but 5% SBC/Rev is almost impossible for SaaS companies (Bentley's 4.8% is an exception – BSY has only 5,500 employees). A realistic convergence point is SBC/Rev ~7-8% → Standard/Owner gap ~8-10x (still significant).

39.2 SBC/Rev × Revenue Growth Two-Variable Matrix

Matrix Interpretation:

• Current Assumptions (12% growth + 10.9% SBC) → ~$220 (approaching the midpoint between Owner PW $193 and Standard PW $238)
• Bull (14%+7%) → $340 (+45%)
• Bear (8%+12%) → $130 (-45%)
• This table illustrates: ADSK's valuation ranges from $130 to $340—a $200 spread entirely determined by two variables: growth rate and SBC

Chapter 40: Product Roadmap → NRR Causal Chain

40.1 ADSK Product Roadmap (2026-2029) and NRR Impact

NRR Causal Chain Diagram:

Net NRR Path Judgment: If Revit Cloud (+2-3pp) + Forma (+1-2pp) + ACC (+1-2pp) are all successful → AECO NRR moves from ~108% → 112-115% → overall NRR moves from ~108% → 110-112%. However, if Neural CAD cannibalizes AutoCAD (-1pp) + Flex consumption model reduces ARPS (-0.5pp) → AutoCAD NRR moves from ~105% → 103-104% → partially offsetting the AECO uplift. Net Effect: Overall NRR fluctuates within the 108-112% range, with the direction depending on the success speed of new AECO features vs. the cannibalization speed of AutoCAD.

Chapter 41: Churn/GRR Segmentation

41.1 Indirect Inference of GRR Segmentation

ADSK does not disclose GRR (Gross Revenue Retention, which only considers churn and not expansion). We use an indirect method to infer:

Inference Method: GRR = NRR - Expansion Rate. If NRR ~108%, Expansion Rate ~8-10% (ARPS increase) → GRR ~98-100%—however, this is an overall average, and segmentation differences can be significant.

Key Findings:

1. F500 GRR ~103% = Net Negative Churn (revenue self-growth) – This is ADSK's most stable foundation. This is due to the triple stickiness of BIM investments by F500 firms: "sunk costs + project lock-in + talent lock-in".
2. SMB GRR ~92-97% = 5-8% annualized churn – Price increases (consecutive price hikes of 7%+ in FY2024-2025) are approaching the payment capacity limits of SMBs. If churn accelerates to >8% → NRR could fall below 100%.
3. Micro GRR ~85-90% = High churn area for AutoCAD LT/Individual versions – These customers are most likely to be attracted by free alternatives (FreeCAD/Zoo.dev).

Causal Chain: Price increases → SMB churn acceleration → Net Adds deceleration (Ch7: from +785K → +516K, -34%) → Growth model shifts from "volume-driven" to "price-driven". The problem is: Price-driven growth has a ceiling (price increases cannot always exceed inflation, otherwise churn will accelerate and offset it). We estimate the price increase ceiling to be CPI + 3-5% (i.e., annual price increases of +5-8%) – exceeding this threshold → SMB GRR sharply declines.

41.2 Implications of GRR for Valuation

If overall GRR decreases from ~98% to ~95% (SMB churn accelerates):

• NRR from ~108% → ~105% (expansion rate unchanged)
• Organic growth rate from ~13% → ~10% (NRR reduction of 3pp ≈ direct 3pp growth rate reduction)
• Impact: Base DCF from $238 → ~$215 (-10%)

This is the underlying mechanism for KS-2 (NRR<100%): NRR falling below 100% means a significant deterioration in GRR → existing customers are churning → growth becomes entirely reliant on new customers (unsustainable) → P/E should be assigned a larger discount.

Chapter 42: Three-Layer TAM Validation

42.1 Top-Down TAM

42.2 Bottom-Up Validation

Validation Conclusion: Top-Down and Bottom-Up are consistent within ±10% – TAM estimate is reasonable. ADSK's SOM is 30% within the AEC SAM ($12B) – an increase in BIM penetration from 50% to 70% could unlock ~$2.4B in incremental SAM (ADSK could capture 60% = $1.4B). This supports AECO's 5Y +12% CAGR (Base Scenario).

Chapter 43: Deep Dive into Five Engines

43.1 Deep Dive into Cyclical Engine: Construction Supercycle + Interest Rate Transmission + Policy-Driven

Construction Supercycle Thesis: Global construction spending from 2024-2028 may be in the midst of the largest "policy-driven supercycle" in 50 years:

• US: IIJA $1.2T Infrastructure Bill + CHIPS Act $280B + IRA $369B → Federal construction spending +50% over the next 5 years
• Europe: Green construction transformation (EU Green Deal) + defense construction (Ukraine reconstruction + NATO infrastructure) → Construction growth +5-8%
• Middle East: NEOM $500B + Saudi Vision 2030 + UAE/Qatar → Middle East construction growth +15-20%
• India: $1.2T Infrastructure Pipeline (NIP 2025) + 100 Smart Cities → India construction +12-15%

Interest Rate Sensitivity Quantification: ADSK revenue elasticity to interest rates is approximately -0.3x (i.e., for every 1% increase in interest rates → ADSK growth rate decreases by 0.3pp). Derivation: (1) Construction spending elasticity to interest rates is approximately -1.5x (interest rates +1% → construction spending -1.5%); (2) ADSK revenue elasticity to construction spending is approximately +0.2x (construction -1% → ADSK revenue -0.2%, because BIM mandates provide a floor). Combined elasticity: -1.5 × 0.2 = -0.3x.

This implies: If the Fed cuts interest rates by 100bps in 2026H2 → construction spending +1.5% → ADSK growth rate +0.3pp. The impact is very small (0.3pp) – confirming the assessment: ADSK is not a pure cyclical stock; BIM mandates make its sensitivity to interest rates significantly lower than that of building materials/construction contractors.

43.2 Deep Dive into Equity Engine: Extreme RSI + Price Action + Technical Pattern

Rarity of RSI 12.45: Looking back at 20 years of ADSK price data, RSI < 15 has only occurred 4 times (2008 Financial Crisis / 2020 COVID / 2022 SaaS Crash / current 2026). 12-month returns after each occurrence:

Statistical Note: An n=4 sample size does not constitute statistical significance (any statistical test P>0.1). However, qualitative observations are valuable: each extreme RSI oversold condition corresponds to "market excessive panic over ADSK" → recovery driven by "panic subsiding + fundamentals materializing." The fundamentals for 2026 (FY2026 Rev +18%, FCF recovery) are better than 2022 (FCF trough 23.3%) → stronger fundamental support for recovery.

Price Position Analysis:

• Current $235 vs 200-day moving average $289 = -18.5% (200-day MA deviation near 5-year high)
• Current $235 vs 52W High $329 = -28.5%
• Current $235 vs 52W Low $215 = +9.5%
• Downside Buffer: $215 is strong support (low point during the most pessimistic period of FY2024 billing transition, tested twice).
• Upside Potential: Return to 200-day moving average $289 = +23% (medium-term target).

43.3 Smart Money Engine Deep Dive: Insider + Institutional + Analyst Three-Dimensional View

Insider Signal Time Series:

Trend: Q1 2024 (after SEC investigation) saw the most concentrated selling (17 transactions); subsequently, both selling frequency and scale decreased — potentially reflecting that "panic selling" after the SEC investigation has ended, but "normalized selling" (routine after RSU vesting) continues. The key is zero buying — if the CEO or CFO makes an initial purchase in the $215-235 range, this would be the strongest Insider bullish signal in the past 5 years.

Analyst Rating Evolution:

Analysts not only refrained from downgrading ratings (Buy% only marginally decreased from 88%→84%) during the 28% share price decline, but also raised their target prices (from $320→$365). This indicates: (1) Analysts believe fundamentals are improving (FY2026 performance is indeed strong); (2) The sell-side has a systematic optimistic bias (unwilling to downgrade due to short-term declines). Our assessment: Analyst target price after a 20% discount ≈ $292 — still implying +24% upside.

43.4 Signal Engine Deep Dive: Strength + Direction + Persistence of 7 Signals

S6 Strengthening cRPO: cRPO (Current Remaining Performance Obligations, revenue to be recognized within 12 months) growth of +23% outpaces reported revenue growth of +18%—this is a **forward-looking positive signal**: it indicates an accelerated accumulation of contracted but unrecognized revenue. cRPO/Revenue = 0.76 (approaching 1.0) suggests that ADSK has approximately 9 months of "revenue visibility." If cRPO growth sustains at 20%+ → FY2027 revenue +12-14% has high certainty.

Counterpoint: The +23% cRPO growth may partly stem from a billing transition (from multi-year contracts → annual contracts → structural increase in cRPO). If the impact of the transition is excluded, organic cRPO growth might only be +15-18%—still positive but smaller in magnitude than the headline figure.

Chapter 44: Operating Leverage Outlook

44.1 OpEx Category Breakdown and Outlook

44.2 Operating Leverage Drivers Breakdown

Largest Driver: S&M Efficiency (-4.9pp)

S&M/Rev has already decreased by 4.1pp from 37.0% (FY2022) → 32.9% (FY2026). Further drivers for decline:

• Direct sales proportion from 37% → 63% → channel commission savings (channel commissions approx. 10-15% vs. direct sales 5-8%)
• FY2026 GTM restructuring (most of the 16% layoffs were S&M-related) → improved human capital efficiency
• Digital marketing replacing traditional sales (BIM products increasingly driven by Product-Led Growth, PLG)