The Deal Gets a Price Tag: Revenue Guide Up, Cash Guide Down, and the Forward Metrics Finally Matter

Results vs. Consensus

Q2 FY27 Scorecard

Year-Over-Year Comparison

Every figure below traces to the condensed consolidated statement of operations in the 8-K Exhibit 99.1 filed August 27, 2026. Percentages are computed from those filed dollars.

Two notes on the comparatives. First, Autodesk reclassified maintenance revenue into subscription revenue beginning in FY27 and restated prior periods; the footnote states the reclassification "did not impact total net revenue", so the total-revenue comparison is clean. Second, the Other-revenue line shows -3% as disclosed by the company; the condensed statement rounds to whole millions ($94M against $96M), so we use the company's rate rather than a rounding-derived one.

Quarter-Over-Quarter Comparison

Q1 FY27 income from operations is derived from the current filing: six-month income from operations of $1,140M less the reported Q2 figure of $599M. It computes to 28.0% of Q1 revenue, consistent with the 28% Autodesk reported at Q1. The most instructive line in this table is the second-to-last pairing: GAAP EPS was flat sequentially at $2.32 to $2.33 while operating income rose 10.7%. The reconciling item is below the line. The 10-Q discloses $60M of upward strategic-investment marks and $15M of negative adjustments across the first half, and by subtraction the upward marks landed in Q1 and the negative adjustments in Q2. Q1 GAAP EPS was flattered by a mark; Q2 was penalized by one. Neither touches the non-GAAP figure, which is why non-GAAP EPS grew 10.4% sequentially while GAAP EPS did nothing.

Quality of Beat

Revenue. $2,046M at +16% reported and +14% constant currency is a genuine beat: above the guide ceiling rather than merely above the midpoint, and broad-based, with AECO at +15% CC, AutoCAD and AutoCAD LT at +11% CC, Manufacturing at +12% CC and Media and Entertainment at +14% CC. Two disclosed flatterers remain and both are shrinking. FX contributed roughly 2 points (the 10-Q's own table shows net revenue at +16% reported against +14% constant currency, with a "Positive" FX impact), and the new transaction model contributed roughly 2 points, down from about 3.5 points in Q1 and heading to about 1.5 points for the full year on management's math. Strip both and the underlying growth rate is close to 12%. That is a healthy number for a business of this size, and it is materially lower than the 16% headline. Investors anchoring on the reported rate will be surprised in FY28 when both tailwinds are gone.

Margins. Non-GAAP operating margin of 41%, up 2 points, is the cleanest line in the print. The GAAP bridge shows where it came from: 42 basis points of gross-margin expansion and 368 basis points of operating-expense leverage, which together account for the full 409 basis points. Marketing and sales, the line the two-year go-to-market reorganization was supposed to fix, grew 10.2% against revenue at 16.1% and fell 160 basis points as a share of revenue. R&D and G&A both also grew below revenue. Crucially, unlike Q1, this quarter's margin expansion is not a restructuring artifact: the restructuring line swung by only $7M year over year, against an $82M swing across the first half. Q1's GAAP margin inflection was substantially the absence of prior-year charges. Q2's is operating leverage. That is the better quality of the two.

EPS. The two EPS figures tell different stories and only one of them should be trusted. GAAP diluted EPS of $2.33, up 59.6%, outran operating income growth of 34.9% because the effective tax rate fell to 17.0% from 31.4%. Applying last year's rate to this quarter's $593M of pre-tax income yields roughly $407M of net income and about $1.93 of diluted EPS on 211M shares, so roughly $0.40 of the $0.87 GAAP increase is a tax-rate effect rather than an operating one. The 10-Q attributes the sub-statutory rate to the interaction of controlled-foreign-corporation tested income and withholding tax against foreign rate differentials, deductible stock compensation, the foreign-derived deduction-eligible income regime and credits. Non-GAAP EPS of $3.30, up 26.0% from $2.62, is the number that reflects the operating quarter: roughly 22 points of it is the non-GAAP operating margin base expanding, and about 2 points is the 1.9% reduction in the diluted share count from the buyback. There is no below-the-line rescue in the non-GAAP figure; the strategic-investment loss is added back and the tax adjustment is a $0.20 subtraction, not a benefit.

Segment Performance

Revenue by Product Family — Q2 FY27

Constant-currency growth decelerated in three of the four named families versus Q1: AECO from 18% to 15%, AutoCAD from 14% to 11%, and Manufacturing from 17% to 12%. That is a consistent 3 to 5 point step-down across the portfolio rather than a single soft spot, which points at the fading new-transaction-model tailwind and at the reorganization's uneven normalization rather than at a demand break in any one product line. It is nonetheless the first quarter in this coverage arc where three of the four named families decelerated at once.

AECO — Still the Engine, Now With the Operations Leg Attached

AECO at $1,029M and +15% constant currency is just over half of total revenue and remains the fastest-growing named family. Management again attributed the strength to construction and emerging markets, and the headline enterprise win of the quarter sits squarely here: a common-data-environment standardization at one of the world's largest retailers that carries through into operations and digital twins, which is precisely the design-to-operate loop the MaintainX acquisition was bought to close.

"With a vision of every project becoming a complete data-rich digital twin, one of the world's largest retailers selected Autodesk Forma as the common data environment connecting planning, design, construction and operations across its portfolio in North America. Rather than optimizing individual phases, it is creating a connected digital thread that carries trusted project information from concept through construction and into operations. This strategy has led the customer to select Autodesk Tandem as its digital twin platform." — Andrew Anagnost, CEO

On the construction sub-business specifically, management held the line on the growth rate and pushed back on reading the uneven construction macro into Autodesk's numbers, arguing that low technology penetration, not backlog oscillation, is the driver.

"So we're super happy with our construction business. It's growing north of 20%... construction is kind of reverting back to the mean in terms of backlogs. We're completely fine with that because the real underlying driver of growth is not the oscillations in the various metrics you're seeing. It's the fact that the penetration of technology in construction is so low right now." — Andrew Anagnost, CEO

Assessment: The technology-penetration argument is the correct frame and has held up through two soft construction years, so we give it weight. The datapoint we would rather have is the AECO number itself: 15% constant currency on the largest base is good, and it is 3 points slower than Q1. AECO is also where the MaintainX attach thesis has to prove itself, and management flagged that MaintainX's installed base is manufacturing-weighted today, meaning the AEC cross-sell is opportunity rather than achievement.

Manufacturing — Fusion Is Working, the Growth Rate Slowed Anyway

MFG at $385M and +12% constant currency is the largest deceleration in the family table, down 5 points from Q1's +17%. The qualitative commentary was, if anything, stronger than last quarter: management described broad customer wins at German industrials and US machine shops and pointed to Fusion as the product where Autodesk's AI work is landing deepest.

"We're seeing not only continued growth in user acquisition, continued growth in ACV and revenue, but we're also seeing continued growth in multi-seat purchases, which is exactly what we want to see... We've been very AI forward with Fusion and not only with regards to the ecosystem, but also with regards to the functionality we develop... This is absolutely a tailwind to Fusion growth, and we expect that tailwind to continue." — Andrew Anagnost, CEO

Assessment: The tension in this segment is that the narrative accelerated while the number decelerated. Manufacturing is the family with the most exposure to a mixed industrial macro and the most exposure to the new-transaction-model tailwind rolling off, so we do not read 12% as a Fusion problem. We do read it as a reminder that Autodesk is now reporting a business whose growth rate is converging toward low double digits from both directions, and that Fusion's qualitative momentum is not yet large enough to hold the segment rate up on its own.

AutoCAD, AutoCAD LT and M&E — The Ballast and the Farewell

AutoCAD and AutoCAD LT at $500M and +11% constant currency is the slowest of the named families and the most important one nobody discusses. It is a very high-margin, very low-churn annuity that funds the platform investment and the buyback, and it is still growing at a low-double-digit rate two decades past maturity. M&E at $92M and +14% constant currency accelerated from +12% at Q1, which is the strongest relative move in the table. It is also the last quarter M&E will be reported as its own family: from Q3 it folds into "Other" alongside the operations business, a change management was pressed on and explicitly denied reading anything strategic into.

Assessment: AutoCAD's 11% is the quiet floor under the whole model, and the reason a 5% free-cash-flow yield on this business is defensible even if growth settles at low double digits. On M&E, we take management at its word that the bucketing is a housekeeping consequence of adding operations rather than a signal, while noting that a segment gets easier to de-emphasize once it stops being separately reported.

Revenue by Product Type — Q2 FY27 (final quarter of this disclosure)

Assessment: Make at +24% constant currency is the single best growth number Autodesk disclosed this quarter, and it is being retired. Management is discontinuing the Design and Make split on the reasoning that convergence makes the buckets meaningless, which is intellectually coherent and commercially convenient: the cut that showed a 24%-growing business inside a 14%-growing company goes away, replaced by qualitative commentary. We think the strategic logic is real and the timing is unfortunate.

Revenue by Geography — Q2 FY27

Constant-currency growth is tightly clustered at 13% to 14% across all three regions, down from a 16% to 17% cluster at Q1. The composition remains the healthiest possible shape, with no single-region dependency and no obvious soft spot, but the level moved down uniformly. EMEA's optically strong +19% reported carries roughly 6 points of FX; the +13% constant-currency figure is the honest read and is also the slowest of the three, consistent with management's disclosure that Western Europe is the last region to normalize on the sales reorganization because of local consultation requirements.

Assessment: Geographic uniformity is a quality marker. It says the deceleration is a company-level tailwind roll-off rather than a regional demand event. The Western Europe lag is the one genuinely operational drag, and management now describes it as a timing issue with green shoots in pipeline rather than an unresolved risk, which is a step forward from Q1's framing.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Management was measured and procedural rather than promotional, a clear step down in temperature from the strategic set-piece of the Q1 call. The prepared remarks led with a customer story rather than a thesis, the guidance discussion was unusually granular about what was organic and what was acquired, and the acquisition itself was handled as an integration project rather than a vision. Where management was least convincing was on the forward metrics: billings, RPO and net retention all decelerated, and each was addressed with the same multi-year-discounting explanation given in the prior two quarters, without a fresh number attached.

1. Billings, RPO and the Deceleration Management Did Not Volunteer

The most important numbers in the release are the ones that did not lead it. Billings grew 10% reported while revenue grew 16%, though on the constant-currency basis this report uses for every other revenue figure the gap is only 2 points (billings +12%, revenue +14%): currency was a tailwind to revenue and a headwind to billings in the same quarter, and it accounts for roughly four of the six reported points. Current RPO decelerated to +12% from +18% one quarter earlier. Total RPO grew 2%, down from 9%. Unbilled deferred revenue fell 8%. And in the 10-Q, filed the next day, net revenue retention was disclosed as having slipped from above the 100% to 110% band a year ago to approximately the top of it. Management's explanation, which is accurate and consistently given, is that Autodesk is deliberately reducing multi-year discounting, which shortens contract durations and depresses unbilled deferred revenue and RPO in exchange for better price realization at renewal.

"During the second quarter, we sustained our program of reducing multiyear discounts established over the last few years, including winding down multiyear renewals for the Maintenance-to-Subscription program. We continue to expect the reduction of discounting for multiyear contracts to benefit price realization over time, while also temporarily weighing on unbilled deferred revenue and RPO growth." — Janesh Moorjani, CFO

The mechanism is real and it does explain total RPO, which is dominated by unbilled deferred revenue and therefore by contract duration. It explains current RPO less well, because current RPO is a twelve-month forward book that is largely insulated from duration mix. Current RPO decelerating 6 points in a single quarter is a demand-adjacent signal, not purely a mix artifact. Management did not address the current-RPO figure specifically, and no analyst pressed on it.

Assessment: This is now the central open question in the Autodesk thesis, and it displaces the sales reorganization as the thing to watch. The bull reading is that FY27 laps an unusually large prior-year renewal cohort and that the discounting programme is a deliberate price-for-optics trade with a payoff in FY29 and beyond. The bear reading is that revenue is outrunning billings even on the cleaner constant-currency basis, that deferred revenue is being harvested, and that FY28 revenue growth converges on the billings rate. We lean toward the first reading because renewals are demonstrably strong and the pricing decision is well-documented, but the margin of safety in that view narrowed this quarter.

2. MaintainX Closed, and It Now Has a Number

The acquisition announced alongside the Q1 print closed on August 3, 2026, three weeks into the current quarter. Guidance now includes it, and for the first time the deal shows up as arithmetic rather than as a strategic argument: roughly $60M of second-half revenue and roughly $70M of second-half billings, both Q4-weighted, against operating losses, financing costs and about $45M of transaction expenses.

"We expect MaintainX to contribute approximately $60 million to second half fiscal '27 revenue and approximately $70 million to second half fiscal '27 billings, both weighted slightly towards the fourth quarter given the growth profile of the business." — Janesh Moorjani, CFO

Integration is deliberately unambitious in the near term. Management described the priority as back-office consolidation and momentum preservation, with the synergy vectors (enterprise-account attach, European expansion, and a push from MaintainX's manufacturing-weighted base into AEC) framed as a multi-year build modelled explicitly on the construction playbook.

"In terms of where we're at with the integration, it's very early. We just closed the acquisition early in August. Our main focus right now is on integrating the back office and maintaining the momentum in the business." — Andrew Anagnost, CEO

Assessment: Restraint is the right posture and it is what the construction precedent actually teaches. The risk in a deal like this is not that synergies arrive slowly; it is that the acquirer breaks the acquired growth engine trying to accelerate them. Management naming momentum preservation as the first-order goal, ahead of cross-sell, is the single most reassuring thing said about the deal on this call. It also means anyone underwriting near-term accretion is underwriting something management has not promised.

3. The Financing Is Disclosed, and It Is the Answer to Last Quarter's Biggest Gap

At Q1 we flagged the absence of debt quantum, rate and pro-forma leverage as the most conspicuous omission on the largest deal in company history. The 10-Q closes that gap in full. Preliminary consideration is approximately $3.53B net of cash acquired, the figure in Note 19 of the 10-Q; the roughly $3.6B cited at announcement and in the filing's own liquidity discussion is the same transaction, rounded. It was funded with three pieces: a $1B unsecured 364-day delayed-draw term loan, arranged in June and drawn on the closing date at a weighted-average rate of 4.58%; a commercial paper programme established July 13 with a $2.0B ceiling, of which $1B was outstanding at July 31 at a weighted-average 4.17%; and cash on hand. Both facilities carry a maximum leverage covenant of 3.50 to 1.00, stepping up to 4.00 to 1.00 for up to four quarters after a material acquisition, and Autodesk was in compliance at quarter end with the $2B revolver entirely undrawn.

The structure explains the July 31 balance sheet, which otherwise looks strange. Cash of $4,098M against $2,249M at January 31 is not a cash-generation story; roughly $1B of it is commercial paper raised in the final weeks of the quarter and warehoused to fund a deal that closed three days after the balance-sheet date. The July 31 balance sheet is the war chest, not the post-deal picture.

Assessment: This is materially better than the vacuum implied. Roughly $2B of the $3.53B is debt-funded at a blended rate near 4.4%, against roughly $2.7B of annual free cash flow, and the covenant headroom is wide. The genuine caveat is tenor rather than quantum: adding the $1B term loan to the $1B of commercial paper and the $500M of notes maturing in FY28 puts roughly $2.5B of a roughly $4.5B post-close debt stack inside twelve months. Autodesk will almost certainly term this out with a bond deal, and the coupon it prints is a real, if second-order, input to the FY28 EPS bridge. A company that was net cash a year ago now has a refinancing calendar.

4. The Q3 EPS Guide Is Financing Arithmetic, Not a Demand Signal

The proximate cause of the sell-off was the Q3 non-GAAP EPS guide of $3.04 to $3.09 against a Street near $3.14. The same guide carries revenue of $2,125M to $2,140M, comfortably above the Street. Revenue above, earnings below, in the same guide, from the same company, for the same quarter. That shape has one obvious explanation.

On our arithmetic, Q3 carries roughly $20M of incremental interest expense against Q2's $23M (a full quarter of $1B of commercial paper at 4.17% plus almost a full quarter of the $1B term loan at 4.58%, against the roughly two weeks of commercial paper already in Q2). It also gives up roughly $20M to $25M of quarterly interest income on the approximately $2.5B of cash deployed at close, against the $28M Q2 earned on a much larger average balance. Combined, that is roughly $40M to $45M per quarter of pre-tax financing drag, or about $0.16 to $0.18 per share after tax at the current rate, before counting MaintainX's own operating losses. The Q3 guide is roughly $0.075 light. The financing drag alone more than covers it.

Assessment: This is the most straightforwardly mispriced element of the print. Autodesk's non-GAAP framework adds back stock compensation, amortisation and acquisition-related costs, but not interest, and management explicitly said the guide includes MaintainX's "operating and net financing costs" from the acquisition. So the deal's financing lands in the headline EPS number the Street compares to its model, and the Street's model largely did not have it. Strip the financing and the implied organic earnings guide is above where consensus sat. That does not make the dilution imaginary. It makes the reaction a category error about which line moved.

5. The Sales Reorganization Got Its Clearest De-Risking Yet

The dominant bear point of the last three quarters was that the two-year go-to-market reorganization completed in January would impair new-business generation. Q1 landed within the range of expectations, which was reassuring but unfalsifiable. Q2 produced the first concrete comparative.

"we saw strength in that in Q2 in most of the areas around the world where they were much closer to plan, and it's a lot better than it was in Q1." — Janesh Moorjani, CFO

Management also sharpened the geographic picture, naming Western Europe as the specific laggard and attributing it to consultation and labour requirements rather than to demand, while describing renewal productivity as a completed win.

"The Americas, APAC, Eastern Europe and the Middle East are normalizing a little earlier than Western Europe. The overall impact to new subscription growth was once again within the range of our expectations, while the linearity of billings during the quarter was better than we expected. Renewal rates remained strong." — Janesh Moorjani, CFO

Assessment: The phrase "a lot better than it was in Q1", measured against seller attainment to plan, is the first management statement on this topic that could have been falsified and was not. It moves the reorganization from an active risk to a residual one. The remaining exposure is concentrated in Western Europe and in the Q4 EBA renewal cohort, both of which management named unprompted. We are downgrading this from the top bear point to the second, and promoting the forward-metric deceleration into its place.

6. Margins: 41% Now, Better in FY28, Still 41% in FY29

The margin framework survived the acquisition intact, which is the commitment we said at Q1 would be the test of whether the deal was being absorbed or accommodated. Non-GAAP operating margin guidance for FY27 is unchanged at approximately 39%, explicitly described as higher underlying margins offset by MaintainX dilution. The FY29 target of 41% was reaffirmed, and management added a forward step it had not previously given.

"We expect fiscal '28 non-GAAP margins to improve modestly from 39% in fiscal '27, with underlying improvement partly offset by the annualization of MaintainX operating costs. We remain on track to achieve 41% non-GAAP operating margin in fiscal '29." — Janesh Moorjani, CFO

Two supporting disclosures matter. Stock-based compensation is guided to about 9% of revenue in FY27, down from about 11% in FY26, which management named alongside underlying margin improvement as the reason GAAP operating margin expanded 4 points. And management confirmed that the AI gross-margin question, which has compressed margins across the software group, is inside the FY29 target rather than a future revision to it.

"We can this year, Brent, and that's what we've factored into the model. Over time, as the AI-based workloads and offerings continue to grow, those will be gross profit dollars accretive but they will put a little bit of compression on gross margin. But that's one of the things that we had also factored into our long-term operating margin target when we rolled that out. So that is embedded in the 41% margin target for FY '29." — Janesh Moorjani, CFO

Assessment: Holding a multi-year margin target through a dilutive acquisition, and volunteering an intermediate FY28 step in the same breath, is a high-conviction signal about the size of the underlying operating leverage. The stock-based compensation walk from 11% to 9% of revenue is the under-discussed engine of the GAAP convergence and is worth more to the through-cycle multiple than most of what was discussed on the call.

7. Autodesk Is Reducing What It Discloses

Three disclosure changes were announced together. Design and Make revenue will no longer be reported. Media and Entertainment folds into an "Other" bucket with operations. MaintainX revenue will be disclosed for four quarters and then presumably not. The stated rationale is that convergence makes the old buckets incoherent.

"And going forward, we will stop disclosing design and make revenue, consistent with our strategy of converging workflows end-to-end and expanding our business into operations. We will continue to provide regular commentary on construction, Fusion and operations and will also disclose MaintainX revenue for 4 quarters to help you track the performance of those businesses." — Janesh Moorjani, CFO

The replacement is qualitative commentary: growth rates spoken on calls rather than dollars printed in releases. Management pointed out, correctly, that the Make bucket never cleanly captured construction anyway.

Assessment: The strategic argument is sound and the timing is poor. Autodesk is retiring the two disclosures that would let an investor independently verify the organic trajectory (the Design and Make split, and eventually the MaintainX line) in the same quarter that its billings, RPO and net-retention metrics all decelerated, and one quarter after taking on $2B of acquisition debt. Each change is individually defensible. Together they mean the FY28 organic-versus-acquired question, which is the central modelling question in the name, will have to be taken substantially on management's word. We do not think this is deliberate obfuscation. We do think it raises the cost of being wrong.

8. Below the Line: Marks, Rates, and Why GAAP EPS Went Nowhere

Two non-operating items distorted the quarter in opposite directions and neither was discussed on the call. First, "Interest and other" swung to a $6M net loss from $12M of income a year ago, and from $58M of income in Q1. The 10-Q shows why: strategic equity investments carried $60M of upward marks and $15M of negative adjustments across the first half, which by subtraction puts the upward marks in Q1 and the writedowns in Q2. Autodesk holds $392M of direct investments in privately held companies, up from $346M at January 31, and marks them through the income statement. Second, the effective tax rate of 17.0% against 31.4% a year ago added roughly $0.40 to GAAP EPS.

Assessment: The net effect is that GAAP EPS is close to uninformative this quarter, moving $0.01 sequentially while operating income rose 10.7%. Both distortions are excluded from the non-GAAP figure, which is one of the rare cases where the non-GAAP number is the more conservative and more honest one. Investors reading the GAAP line in isolation got a flat quarter; investors reading operating income got a good one. The tax rate is the item to watch: a sustained sub-20% effective rate is worth real money to GAAP earnings and to the cash tax line, and nothing in the disclosure suggests it is a one-quarter event.

9. AI: Orchestration, Not a Revenue Line

The AI section of the prepared remarks was the most developed part of the call and the most explicitly positioned against the horizontal-model competitive threat. Management's argument is that the winning position is orchestration across models rather than ownership of the best single model, with Autodesk's own NeuralCAD used where geometric precision matters and frontier models used where they do not.

"The future of AI won't belong to the company with the best single model. It will belong to the platform that combines the richest context with the right models to deliver the best outcomes for customers." — Andrew Anagnost, CEO

On monetisation, management again declined to break out an AI revenue figure, arguing that task-level automation is currently a subscription accelerant rather than a separate product, with consumption-based revenue arriving later as workflow and system-level automation ship.

"I think you should look at the task-based automation as not only a lubricant, but an accelerant to some of the subscription-based businesses as we move forward, and we'll treat it as such. But as time goes on and we move deeper into workflow automation, systems level automation, you're going to see us start to express more consumptive revenue." — Andrew Anagnost, CEO

Assessment: The refusal to manufacture an AI revenue number is defensible and, in a group where those disclosures have become marketing artefacts, arguably a mark of discipline. The cost is that the AI leg of the thesis remains unmeasurable for a fourth consecutive quarter. Autodesk University in a few weeks is now the third consecutive event flagged as the place where specifics arrive; that is the last time we will treat the deferral as neutral rather than as evidence.

10. Consumption Pricing Is Arriving Through the Side Door

Two smaller commercial moves point at the same destination. Autodesk lowered the Flex minimum purchase in early June to reach smaller customers, a change put to management on the call as a cut from $300 to $99 and not disputed, and it is proceeding with charging for API and machine-driven usage despite pushing some timelines to accommodate customer feedback.

"Customers are very much aware that we're going to be charging for APIs and machine-based usage. We've actually already charged for some of that usage. We do listen to customers. We do adjust some of our time lines and our intent to make sure that we're matching with customer expectations and customer ramp-up times, but continue to look for us to charge for access to machine-driven usage to our capabilities, our IP and our products. That's not going to change." — Andrew Anagnost, CEO

On Flex, management reported early adoption gains and, notably, that the change is accretive rather than dilutive.

"So yes, we have seen some acceleration in adoption of Flex, and it has been net accretive to the business, which is great." — Andrew Anagnost, CEO

Assessment: Flex is the vehicle through which consumption revenue will eventually show up, and management said as much: as workflow automation deepens, consumptive revenue "will show up through Flex business and other things". That makes the $99 minimum more strategically significant than a pricing tweak, because it seeds the token-based motion in the long tail before the AI workloads that need it exist. The API charging posture, stated as flatly as management has ever stated it, is the clearest evidence yet that agentic access will be a priced surface rather than a free one.

11. Leadership: The AECO Succession

Amy Bunszel, EVP of Architecture, Engineering and Construction Solutions, announced her retirement after 23 years. She runs the segment that is just over half of revenue, that carries the construction growth engine, and that is the primary attach surface for MaintainX. She remains through the search and transition.

"Amy has helped shape Autodesk into the company we are today. Her vision, deep commitment to our customers and leadership through some of our most important transformations have built a world-class AEC business and an enduring legacy." — Andrew Anagnost, CEO

Assessment: A planned retirement with an open-ended transition is the low-risk version of this event, and no successor has been named. The reason it belongs in a recap rather than a footnote is timing: AECO is where the operations strategy has to be executed, and the executive who built the segment is leaving in the quarter the operations acquisition closed. We do not read it as a warning signal. We do add the succession to the watch list, because an external hire into AECO would be a meaningfully different signal than an internal promotion.

Guidance & Outlook

Guidance now includes approximately six months of MaintainX. The CFO framed the raise as a combination of the Q2 outperformance, a genuine improvement in second-half expectations, and the acquisition, with an FX and transaction-model-mix offset working the other way.

"We have increased our fiscal 27 billings and revenue growth guidance to reflect higher underlying growth expectations, as well as the incremental contribution from MaintainX. Our fiscal 27 non-GAAP margin guidance is unchanged, reflecting higher underlying margins from operating leverage and benefits from our go-to-market optimization, offset by margin dilution from the MaintainX acquisition." — Janesh Moorjani, CFO

Decomposing the revenue raise. The FY27 revenue guide midpoint moved from $8,185M to $8,320M, a $135M increase. MaintainX supplies $60M of that. The Q2 beat against the guide midpoint supplies another $36M. The residual, roughly $39M, is the genuine second-half organic raise. That is modest in dollars but real, and it is the number that matters, because it says management raised its own operating expectations on top of flowing through a beat.

Implied second-half ramp. Against the FY27 revenue midpoint of $8,320M and first-half actuals of $3,980M, the implied second half is $4,340M. Netting the Q3 guide midpoint of $2,132.5M leaves an implied Q4 of roughly $2,207.5M. Against FY26 comparatives, that is Q3 growth near 15% and Q4 growth near 13%, versus 17.2% in the first half. Growth steps down through the year by design, as the transaction-model tailwind fades from roughly 3.5 points in Q1 to roughly 2 points in Q2 to roughly 1.5 points for the year. Excluding MaintainX entirely, the implied second half is about $4,280M, or roughly 12% growth. That 12% is the cleanest available read on the underlying rate heading into FY28.

Cross-check on the EPS guide. First-half non-GAAP EPS was approximately $6.29 ($2.99 in Q1 plus $3.30 in Q2). Against the FY27 midpoint of $12.56, the implied second half is roughly $6.27, and netting the Q3 guide midpoint of $3.065 leaves an implied Q4 near $3.21. The full-year guide is arithmetically consistent with what the first half actually delivered, with no back-half hockey stick embedded.

Street positioning. The revenue guide sits above consensus at every horizon: Q3 at $2,125M to $2,140M against a Street in the low $2.08B area, and FY27 at $8,295M to $8,345M against roughly $8.21B. The EPS guide sits below at both: Q3 at $3.04 to $3.09 against roughly $3.14, and FY27 at a $12.56 midpoint against roughly $12.60. Free cash flow narrowed with the midpoint down roughly $25M, absorbing about $45M of MaintainX transaction expenses plus net financing costs. The divergence between the revenue and earnings lines is the entire story of the guide.

Guidance style. Conservative, and explicitly so. The CFO described the philosophy as unchanged and named the two remaining sources of variability: the Q4-concentrated EBA renewal cohort, which is FY27's largest, and the pace of new-business normalization in Western Europe. Six months of visibility has replaced twelve, which naturally narrows the range, and management said the raise reflects both better visibility and the work still ahead. On the pattern of the last several quarters, the FY27 range looks more likely to be walked up again than defended.

Analyst Q&A Highlights

Growth Durability Once the FY27 Tailwinds Lap

The single most thesis-relevant exchange of the call went straight at the question the deceleration raises: FY27's reported growth benefits from the transaction-model change, the completed shift to annual billings, and unusually large renewal cohorts, all of which reverse or lap in FY28. The question asked whether those become an outright headwind. Management declined to guide but enumerated the moving parts candidly, confirming both the transaction-model roll-off and the full-year MaintainX cost load, and then pivoted to the durability of the underlying renewal and enterprise-agreement base.

Q: "So you've done a nice job sustaining a consistent level of growth. But as we look beyond FY '27, Autodesk is going to start to lap a few tailwinds related to the model change and the transition to annual billings and a couple of the larger renewal cohorts that you talked about, which have all been beneficial to underlying growth. So as we think about that, is it possible that, that at some point starts to become a headwind to billings growth? And if not, maybe you could just talk about the incremental growth drivers that you're excited about that could help support the growth durability that we've seen?"
— Taylor McGinnis, UBS

A: "So as we think about fiscal '28, there's a number of different moving parts, as you mentioned, and it's a little bit early to talk about '28, but I'll at least just highlight some of those moving parts for you. First off, as you know, fiscal '27 has some benefit from the new transaction model that will not recur in fiscal '28. So that's one factor to consider. And '28 will also include the full year contribution from MaintainX, as I just mentioned, both in terms of the top line as well as the expenses. We'll also have the continued normalization of new business productivity where -- as we continue to get better in Western Europe and the rest of the world, that will ultimately be a driver of the business for us."
— Janesh Moorjani, CFO

Assessment: Management neither denied the headwind nor quantified it, which is the correct posture six months out but leaves the FY28 bridge to be built by investors. The answer implicitly confirms the frame: FY28 loses about 1.5 points of transaction-model tailwind, carries a full year of acquisition costs, and depends on Western European productivity recovery to offset both. Notably, the answer's list of positive drivers was qualitative (renewals strong, EBAs strong, demand sustained) while the list of headwinds was specific. That asymmetry is a fair summary of where the FY28 setup stands.

How Much of the Raise Was Bought and How Much Was Earned

The most immediately useful exchange was a request to separate the organic raise from the acquired one, prompted by investor inbounds during the call. Management gave a precise decomposition on billings, walking the prior organic guide range to the new one and isolating the acquisition's contribution as a single percentage point.

Q: "I think we're getting some inbounds on how much exactly was the revenue and the billings raised for the full year inorganically versus the organic contribution. I know that you gave us the inorganic contribution dollars in the press release, but can you just help us fully clarify the math on just by how much organically you raised revenue and billings for the full year?"
— Joshua Tilton, Wolfe Research

A: "The underlying billings outlook improved by about 2 percentage points. 1 percentage point of that comes from MaintainX, which is approximately $70 million and roughly 1 percentage point comes from the stronger underlying performance that we had in the business. There's an offset on that, which is we had a little bit of headwind from currency and transactional model mix. And so that offsets about 1 percentage point... when you think about our prior guidance, excluding the new transaction model and excluding currency, that prior guide was 8% to 9% year-over-year growth. Now that has increased to 9% to 10% year-over-year growth on an organic basis and tack on 1 percentage point for MaintainX, so that's 10% to 11% year-over-year growth."
— Janesh Moorjani, CFO

Assessment: This is the most valuable number given on the call and it deserves to be the anchor for FY28 modelling. Organic billings growth ex-transaction-model and ex-currency was raised from 8% to 9% up to 9% to 10%. That is a real raise on the cleanest available basis, and it corroborates the roughly 12% underlying revenue rate we derive independently from the second-half guide. It also sets the ceiling: an organic business compounding billings at 9% to 10% is a good business, and it is not a mid-teens grower once the optics fade.

How Much Prudence Is Left in the Back-Half Guide

A question on guidance philosophy asked how much conservatism remains embedded now that only six months are being forecast. Management's answer was more forthcoming than usual, naming the specific unresolved items rather than gesturing at prudence generically, and pairing genuine confidence with two named risks.

Q: "You talked a little bit about guidance philosophy in your prepared remarks. I was just wondering if you could talk a little bit about maybe how we should think about the levels of prudence built into the Q3 and full year guide?"
— Saket Kalia, Barclays

A: "at the start of the year, our guidance had to contemplate a much wider range of outcomes, particularly around the impact of the sales reorganization on new business productivity. And 6 months into the year, the impact from the reorg has largely remained within the range we expected. Renewals have been very strong. The business has performed much better than our initial assumptions in many areas... There is naturally less uncertainty today compared to the past because we're only forecasting the remaining 6 months of the year. But at the same time, we've also got work ahead of us in the second half. We have our largest EBA renewal cohort, as you know, with a concentration in Q4. And we're also focused on the pace of new business productivity normalization, particularly in Western Europe."
— Janesh Moorjani, CFO

Assessment: "The business has performed much better than our initial assumptions in many areas" is a stronger statement than the guide's modest organic raise implies, and it is the kind of remark that usually precedes further upward revisions. The two named risks are both concentrated in Q4, which means the FY27 outcome is more back-end loaded than the smooth guidance range suggests. The Q4 EBA cohort is the swing factor for the year and it will not be resolvable until the February print.

Leading Indicators on Sales-Force Productivity

A question pressed for the measurable evidence behind the claim that reorganization disruption is receding, asking specifically which indicators improved and which have not yet turned. The answer separated the completed win (renewal productivity, the original objective) from the work in progress (expansion and new business, with Western Europe explicitly the laggard), and then supplied the concrete comparative the prior quarter had lacked.

Q: "And then just as a follow-up, on this sales reorganization, what are some of the leading indicators that you're seeing now in terms of rep productivity, pipeline creation or some of the time to ramp that improved in Q2? And maybe what are some of the ones that you're still looking to turn the corner? And how are you looking at signs just to confirm that the disruption has weakened more of that risk is actually behind us now?"
— Elizabeth Porter, Morgan Stanley

A: "One of the places we've seen actual real productivity increases, and it was one of the places we were targeting is the renewal business. Our renewals are much more productive now, both internally and through the channel partners... We've also seen all the green shoots associated with the expansion business internally and with the channel. The one area that Janesh highlighted is, and this is because of consultation periods and all the things associated with mature Europe, it's been a little slower to ramp up on the expansion business. We're seeing the signs. We're seeing continued growth in pipeline associated with kind of forward-looking indicators that tell us we're moving in the right direction."
— Andrew Anagnost, CEO

Assessment: The valuable half of this exchange came in the CFO's follow-on, which defined the metric as seller attainment to plan and stated that in Q2 sellers were "much closer to plan, and it's a lot better than it was in Q1". That is a falsifiable comparative on a specific internal measure, and it is the strongest de-risking datapoint produced in three quarters of this topic. The residual exposure is now geographically bounded rather than company-wide, which is a different and smaller risk than the one carried into this print.

What MaintainX Actually Does to the Margin Walk

A question asked, for the first time on the record, what MaintainX's standalone margin profile looked like before the acquisition and what the drag will be. Management answered the profitability question directly and declined the quantification, while restating the FY28 direction of travel.

Q: "I know you reiterated your operating margin guide. But now that MaintainX has closed, curious what might be the operating margin impact in the quarter of the year or maybe historically what the margin profile looked like before you acquired it?"
— Jason Celino, KeyBanc Capital Markets

A: "MaintainX is a high-growth business when we acquired it and still is. And consistent with many high-growth businesses, as you might imagine, the business was not profitable. And so as we bring MaintainX into Autodesk, our main focus is to ensure that we do everything that we can to continue to preserve the strong growth rate and continue to fuel the engine. So MaintainX does present a drag on operating margin in fiscal '27. And then next year, we'll have the full year annual cost of MaintainX from an operating perspective and also in terms of the net financing costs. But despite that, as I mentioned in my prepared remarks, as I think about next year, we do expect that we will increase the operating margin next year from the 39% that we have this year."
— Janesh Moorjani, CFO

Assessment: The confirmation that MaintainX was unprofitable at acquisition is new and it is the correct thing to have said. It also sharpens the tension we flagged at Q1: management has committed to preserving MaintainX's growth rate and to expanding consolidated margin in the same year, and those two goals compete for the same dollars. The commitment to raise the FY28 margin above 39% despite a full year of operating and financing costs is the more demanding of the two promises and the one to grade next February.

Why the Disclosure Changes, and Why Now

The disclosure reductions drew the sharpest challenge of the call: why change mid-year, and what specifically replaces the retired metrics. Management grounded the timing in the acquisition and committed to ongoing verbal commentary on the three businesses investors care about, plus a time-boxed MaintainX revenue disclosure.

Q: "Maybe just at a high level, like with these disclosure changes, Janesh, I guess, why now? Why sort of midway through the year? And I think there's a lot of good momentum in the construction business. You talked about a leading organization choosing Autodesk Construction this quarter. So can you just be a little bit more specific on how you'll be updating us on the progress of that and the metrics that you'll share?"
— Tyler Radke, Citi

A: "In terms of the why now, it's really got to do with the launch of the operations business, right? With closing of MaintainX, we needed to revisit our product family reporting... construction is reported today under the make business, but even that doesn't entirely capture all the construction business. And so it's really important for us to make sure that as we move forward, we are providing you with a view on how the underlying businesses are performing. So we will provide you with regular updates on that business. It will be regular commentary. Just like we have today, we will tell you how the business has been growing. It's growing north of 20% and has done for some time."
— Janesh Moorjani, CFO

Assessment: The answer is honest about the mechanism and evasive about the substitution. "Regular commentary" spoken on a call is not the same instrument as a reported dollar figure in a press release, and the difference matters most precisely when the underlying trend is decelerating. The construction growth rate stated here, above 20% and sustained, is genuinely good news buried in a defensive answer. We would rather have it printed than said.

Why There Is Still No AI Revenue Number

With software peers increasingly publishing AI revenue disclosures, the call asked whether Autodesk intends to follow. Management declined and reframed around its three-stage monetisation model, and the follow-up on gross margin drew the more consequential answer of the pair.

Q: "Many of your software peers are unveiling AI revenue. I don't know if it's important to you, maybe it's found in other ways. But how do you think about the impact? And is this something you expect that you want to talk about going forward?"
— Brent Thill, Jefferies

A: "Right now, what we're doing is we're enabling lots of task-based automation in the product, lots of AI-driven automation. It's basically built into some of the subscription offerings we have. And like I've said, it's actually a tailwind to Fusion growth right now... But as time goes on and we move deeper into workflow automation, systems level automation, you're going to see us start to express more consumptive revenue. That will show up through Flex business and other things that are also embedded in how we're reporting in our industries. So look for us to evolve these things over time. We're not trying to do one-off presentations on some of these things."
— Andrew Anagnost, CEO

Assessment: "We're not trying to do one-off presentations on some of these things" is a deliberate swipe at the disclosure practices of the peer group, and there is a real argument that a manufactured AI revenue line is worse than none. The problem is that the AI leg of the bull case has now been unmeasurable for four quarters running while being cited as a competitive moat. The follow-up answer was more valuable: AI-driven gross-margin compression is already inside the FY29 41% target rather than a future revision to it, which removes one of the more plausible ways the margin framework could have broken.

What They're NOT Saying

1. Current RPO decelerating 6 points. Current RPO went from +18% to +12% in a single quarter and neither management nor any analyst addressed it directly. The multi-year-discounting explanation covers total RPO and unbilled deferred revenue, which are duration-sensitive. It does not cleanly cover a twelve-month forward book. This is the largest unexplained number in the release.
2. Net revenue retention slipped, and only the 10-Q says so. NR3 moved from "slightly above" the 100% to 110% band a year ago to a level that "approximated the top end" of it. The change is small and the level is still healthy. It was disclosed in a filing the day after the call and was not mentioned in the prepared remarks, which is a notable choice for a metric management describes as helping to monitor "the long-term health of our business".
3. The refinancing plan for $2.5B of near-term maturities. The 10-Q discloses $1B of 364-day term loan, $1B of commercial paper and $500M of notes all maturing within twelve months. Management said nothing about how or when this gets termed out. The coupon Autodesk prints on a replacement bond is a direct input to FY28 EPS and is currently unmodellable.
4. MaintainX margin dilution in basis points. Management confirmed the business was unprofitable at acquisition and that it drags FY27 margin, but gave no figure for the drag, no revenue-growth expectation post-close, and no timeline to profitability. The four-quarter revenue disclosure will show the top line and nothing beneath it.
5. Why the effective tax rate halved. The 10-Q lists the structural drivers of the sub-statutory rate but does not explain the year-over-year move from 31.4% to 17.0%, which is worth roughly $0.40 of quarterly GAAP EPS. Nothing in the disclosure indicates whether the FY27 rate is the new normal or a temporary interaction.
6. A named successor for AECO. The executive running just over half of revenue announced her retirement on the call, with the search still open and no internal or external candidate indicated.

Market Reaction

• Pre-print setup: ADSK closed at $270.58 on August 27, the print day, down 8.6% year to date against the S&P 500 up 12.9%, down 5.4% over the trailing twelve months, but up 14.0% over the trailing thirty days. The 52-week closing range entering the print was $187.72 to $326.79.
• The print-day run-up: The stock rallied 6.2% on August 27 itself, from $254.77 the prior session, on 3.96M shares against roughly 1.7M across the preceding four sessions, about 2.4 times. That entire move happened before any results were released.
• Reaction session: ADSK opened August 28 at $261.16, a 3.5% gap down, traded a $253.16 to $267.00 range, and closed at $260.66, down 3.7% or $9.92 on the day. Volume was 3.4M against a 2.0M thirty-day average, 1.7 times normal. The S&P 500 was down 0.2% on the same session.
• Round trip: Measured from the August 26 close of $254.77, the stock finished the reaction session at $260.66, up 2.3% across the print. Through the August 31 close of $258.53, the last completed session before publication, it sits 4.5% below the pre-print close and 1.5% above where it traded the day before the print.

The headline that ADSK fell on a beat is arithmetically true and framing-dependent. The stock gained 6.2% into the print on heavy volume, gave back 3.7% on the results, and still finished the reaction session above where it started the week. What actually happened is that a run-up got unwound, not that a business was repriced.

Within the reaction, the mechanism was specific and consistent across coverage: the Q3 non-GAAP EPS guide landed below the Street, and the free-cash-flow range was narrowed with the top end cut. Both are direct consequences of the acquisition closing. The revenue guide, which moved up at every horizon and above consensus at every horizon, did not carry the day. That is a predictable outcome when a company changes its capital structure between one guide and the next: models get updated on the line that moved down before the line that moved up.

The session after the reaction added little: August 31 closed at $258.53, another 0.8% lower on lighter volume. That is drift rather than a second verdict, and we would not read a thesis into it. What it does confirm is that the print produced no re-rating in either direction, which is consistent with our own read that the forward metrics, not the guide, are the substantive issue here, and it is why we are maintaining rather than raising our conviction.

Street Perspective

Debate: Is the Billings and RPO Deceleration Mix or Demand?

Bull view: The deceleration is fully explained by a deliberate, disclosed, multi-quarter decision to reduce multi-year discounting, which shortens durations and mechanically compresses unbilled deferred revenue and total RPO in exchange for better price realisation at renewal. Renewal rates are strong, recurring revenue is 97% of the total, and organic billings growth ex-transaction-model and ex-currency was raised from 8% to 9% up to 9% to 10%. A company deliberately trading reported RPO for future pricing is doing the right thing and being punished for the optics.

Bear view: Revenue is outrunning billings on both bases, which means deferred revenue is being harvested and the gap has to close. Current RPO, which is duration-insensitive, still decelerated 6 points in a quarter, and net revenue retention slipped in the same period. The discounting explanation has now been given three quarters running without a quantification of how much of the gap it accounts for. Meanwhile the disclosures that would let investors check the answer are being retired.

Our take: The bulls have the better of the total-RPO argument and the bears have the better of the current-RPO one. Our reading is that most of the deceleration is genuine mix and cohort timing, and that some of it is the underlying rate settling into low double digits as the transaction-model tailwind exhausts. Both can be true, and the combination still supports the rating. The reason this is now our top watch item rather than a resolved question is that management chose not to engage with the current-RPO number specifically.

Debate: Did the Market Misread the Q3 Guide?

Bull view: The guide raised revenue above consensus at every horizon and lowered EPS by less than the financing cost of a deal that closed three weeks into the quarter. On any reasonable estimate of the incremental interest expense and forgone interest income, the financing drag exceeds the $0.075 Q3 shortfall on its own, meaning the implied organic earnings guide went up. The sell-off is a modelling lag, not a judgement.

Bear view: The dilution is real regardless of its cause. Autodesk is paying about 4.4% on roughly $2B of short-dated debt, has given up interest income on roughly $2.5B of deployed cash, absorbed $45M of transaction expenses, and acquired an unprofitable business, all to add $60M of second-half revenue. The free-cash-flow guide came down. Calling that a misread does not make the cash reappear.

Our take: The bull case is correct on the mechanics and the bear case is correct that the mechanics do not matter to a cash-flow yield. Our resolution is that the near-term dilution is a known, bounded, one-year phenomenon whose cost is now fully disclosed, and the acquired asset compounds against it. The important consequence is not the direction of the FY27 EPS guide but the fact that FY28 carries a full year of this, which is why we build our FY28 estimate conservatively rather than extrapolating the FY27 margin walk.

Debate: Does a 21x Multiple With Reduced Disclosure Compensate for the Uncertainty?

Bull view: At roughly 21 times the FY27 non-GAAP EPS guide midpoint and a free-cash-flow yield near 5%, ADSK trades meaningfully cheaper than at the Q1 print, on a 97%-recurring-revenue business with net retention around 110%, 41% non-GAAP margins, a reaffirmed path to 41% in FY29, a de-risked go-to-market reorganization, and roughly $1.4B of annual buyback. Multiple compression has already priced a good deal of the growth normalisation.

Bear view: The multiple is reasonable for a mid-teens grower and merely fair for a low-double-digit one, and FY28 looks like the latter. On top of that, the two disclosures that would allow independent verification of the organic trajectory are being withdrawn, the acquisition is unproven, half the debt stack needs refinancing, and the AI leg of the story remains unquantified after four quarters of promises. Paying a fair multiple for reduced visibility is not a bargain.

Our take: We side with the bulls on the arithmetic and take the bears' point on visibility seriously enough to reduce conviction rather than the rating. A 5% free-cash-flow yield on a franchise with this recurring-revenue quality and this margin trajectory is genuinely attractive, and the pullback has done real work on the entry price. But the disclosure reductions raise the cost of being wrong, and that argues for maintaining a position rather than pressing one.

Model Update & Valuation Framework

Valuation framework. At the $260.66 reaction-session close and the FY27 non-GAAP EPS guide midpoint of $12.56, ADSK trades at approximately 20.8 times forward earnings, against roughly 23 to 24 times after the Q1 print. On the FY27 free-cash-flow guide midpoint of $2,737.5M and 211M weighted-average diluted shares, free cash flow per share is approximately $12.97, a yield of approximately 5.0% at that level, and the same at the August 31 close of $258.53. For a business with 97% recurring revenue, net retention near 110%, a 41% non-GAAP operating margin and a reaffirmed path to 41% in FY29, that is a reasonable price rather than a demanding one.

How we get to FY28. Our $14.00 to $14.30 estimate assumes revenue growth of 9% to 11%, anchored on management's organic billings framework of 9% to 10% plus the MaintainX annualisation and net of the roughly 1.5 points of transaction-model tailwind that does not recur; a non-GAAP operating margin near 39.5%, at the low end of "improve modestly from 39%"; a full year of financing costs on the current debt stack with no assumption of a favourable refinancing; and roughly 2.5% of share-count reduction from a maintained $1.4B buyback. That is deliberately conservative on all four inputs. It cross-checks against the FY27 base: first-half non-GAAP EPS of approximately $6.29 against a full-year midpoint of $12.56 implies no second-half acceleration is required, so FY28 starts from a clean rather than a stretched base.

Revised risk-reward. Against the $260.66 reaction close, the base case of $300 to $340 implies approximately +15% to +30%; the bull case of $350 to $385 implies approximately +34% to +48%; the bear case of $215 to $240 implies approximately -18% to -8%. The base case corresponds to roughly 21 to 24 times our FY28 estimate, the bull case to roughly 24 to 27 times, and the bear case to roughly 15 to 17 times. Base-to-bear upside-to-downside is roughly 1.8 to 1 at the midpoints, narrower than the roughly 3 to 1 we carried at Q1. That narrowing is deliberate: we have lowered the base case to reflect a lower underlying growth rate, not to reflect a lower price.

Thesis Scorecard Post-Earnings

Overall: The thesis is intact but its composition shifted materially. Two bull pillars were confirmed (margins, free cash flow), one was confirmed by the deal closing (platform expansion), and two were challenged (revenue compounding, analytical clarity). On the bear side, the point that dominated the last three quarters was meaningfully de-risked, the leverage point was quantified and contained, and a new one was activated: the forward metrics decelerated across billings, current RPO and net retention simultaneously. Net, the operating quarter strengthened the thesis and the forward disclosure weakened our visibility into it. Those roughly offset, which is why the rating holds and the conviction does not.

Action: Hold existing positions and add selectively on further weakness. The pullback has taken the multiple from roughly 23 to 24 times forward earnings at the Q1 print to roughly 21 times, which is a fair entry rather than a compelling one. We would want either a re-acceleration in current RPO or the first quantified AI monetisation datapoint before adding aggressively.

Bottom Line: A Good Quarter, a Cheaper Stock, and a Narrower Margin of Safety

Rating decision. We maintain Outperform on ADSK. The quarter cleared every near-term bar: revenue and non-GAAP EPS both above the guide ceiling and above the Street, 41% non-GAAP operating margin, 409 basis points of GAAP operating-margin expansion of which roughly 90% was operating-expense leverage, free cash flow up 24%, and a full-year revenue guide raised above consensus with a genuine organic component on top of the acquisition. The bear point that has dominated this coverage arc for three quarters, that the go-to-market reorganization would impair new business, received its first falsifiable de-risking. The stock is roughly 10% cheaper on forward earnings than it was after the Q1 print for a business whose margin framework is unchanged and whose FY28 margin was just guided higher.

Why the market's reaction was the wrong shape. Autodesk raised its FY27 revenue guide midpoint by 1.6% and its non-GAAP EPS guide midpoint by 0.3%, and narrowed free cash flow with the top end cut. Read together, that is a company telling investors precisely what its acquisition costs. The Q3 EPS guide came in roughly $0.075 below the Street, and the incremental interest expense plus forgone interest income from a deal that closed three weeks into the quarter is, on our arithmetic, worth more than that on its own. The operating guide underneath went up. A stock that rallied 6.2% into the print and gave back 3.7% on it did not get repriced; a run-up got unwound.

Why we are not upgrading. Billings grew 12% in constant currency against revenue at 14%, a narrower gap than the reported 10-against-16 suggests but still the wrong way round. The sharper signal is current RPO, which decelerated to 12% from 18% in a single quarter, and that measure is not duration-sensitive, so the multi-year-discounting explanation covers it least well of any metric it was offered for. Net revenue retention eased, disclosed only in the 10-Q. Constant-currency growth decelerated in three of four named product families and in all three geographies together. And in the same quarter, Autodesk announced it will stop disclosing Design and Make revenue, fold Media and Entertainment into an "Other" bucket, and report MaintainX revenue for four quarters only. Each change has a coherent rationale. Together they mean the central FY28 modelling question, how much of the growth is organic, will increasingly be answered by management commentary rather than by reported figures, at the moment that question is hardest.

What would change our mind.

• Downgrade to Hold: a second consecutive quarter of current-RPO deceleration without a quantified explanation; organic billings growth guided below 9% for FY28; a refinancing of the near-term maturities at a materially higher cost; or evidence that MaintainX's growth rate has broken post-close.
• Downgrade to Underperform: underlying constant-currency growth falling below high single digits, a cut to the FY29 margin framework, or a buyback reduction to service the debt.
• Upgrade conviction: current RPO re-accelerating toward mid-teens, a quantified AI monetisation disclosure at Autodesk University, or FY28 guidance above 10% organic revenue growth with the margin walk intact.