Autodesk's 16% Growth Rests on a Shrinking Backlog

Autodesk delivered a beat-and-raise on Aug. 27, and the market applauded. Second-quarter revenue came in at $2.05 billion, up 16%, ahead of both the company's own guidance and the roughly $2.01 billion Wall Street consensus; non-GAAP earnings of $3.30 a share beat the $3.12 estimate by nearly 6%; and full-year revenue guidance went up for the second straight quarter, to $8.30–$8.35 billion. The stock closed up about 6% and has now rallied roughly 15% in three weeks.

The same press release carried a quieter set of numbers that points the other way. Billings grew only 10%; total remaining performance obligations, or RPO, the backlog of contracted revenue, grew 2%; unbilled deferred revenue, work contracted but not yet on the balance sheet, fell 8%. Revenue can grow 16% while the order book grows 2% for only so long before one of those numbers starts moving to meet the other. The duel is over which one moves, and the stock has already rallied on the bull's answer.

Shared facts

• As of Aug. 28, ADSK trades at $270.58, up 6.2% on the report, with a 52-week range of $185.50 to $329.09. Market value is about $57 billion on roughly 211 million shares, roughly 6.9 times this year's guided revenue, about 21 times this year's guided non-GAAP EPS, and about 21 times guided free cash flow.
• Q2 FY2027 (quarter ended Jul. 31): revenue $2,046 million (+16% reported, 14% constant currency); non-GAAP EPS $3.30; GAAP EPS $2.33; billings $1,854 million (+10%); free cash flow $561 million (+24%); non-GAAP operating margin 41%, up two points.
• Backlog as of Jul. 31: current RPO (next-twelve-months revenue) $5,245 million, up 12%; total RPO $7,433 million, up 2%; unbilled deferred $3,175 million, down 8%; deferred revenue $4,258 million, up 11%.
• Fiscal 2027 guidance, raised: revenue $8,295–$8,345 million; non-GAAP EPS $12.52–$12.60; billings $8,575–$8,650 million; free cash flow $2,725–$2,750 million. Third-quarter guide: revenue $2,125–$2,140 million, non-GAAP EPS $3.04–$3.09, now including the acquisition.
• The $3.6 billion all-cash MaintainX purchase closed Aug. 3, a company barely 2% of Autodesk's size bought for roughly 30 times trailing revenue, partly financed with debt.
• The year before: fiscal 2026 (ended Jan. 31) produced revenue of $7.206 billion (+18%), billings of $7.771 billion (+30%), free cash flow of $2.409 billion (+54%), and non-GAAP EPS of $10.43.

Round 1 — Where did the 16% come from?

The bull's cleanest claim is that the gap between revenue and billings is the point, not the problem. For two years Autodesk has been unwinding its old sales-led channel: cutting multi-year discounts, pushing customers to annual deals, and ending multi-year maintenance-to-subscription renewals. Management attributes the shrinking backlog, in its own words, to "a sustained program reducing multi-year discounts." Stop selling tomorrow's revenue at a discount, collect cash every year, book it sooner, and margins lift. Current RPO is still up 12%, deferred revenue is up 11%, the non-GAAP margin hit 41%, and free cash flow rose 24% in the quarter. The backlog is leaner because the company stopped filling it with cheap future revenue. That is a quality upgrade, not a demand failure.

The bear's answer is that the forward numbers are the headline and the P&L is the echo. Billings, the closest thing to new orders, grew 10%. Total RPO grew 2%. In the first quarter, 3.5 points of the 18% growth came from the transaction model. Add the near-45% cumulative price increases Autodesk has pushed through since 2016, and the 16% is a stack of mechanics — model timing plus price — rather than new customers at the same pace. The channel disruption is real: a European reseller's design business fell 12% organically as Autodesk rewired partner incentives. When the backlog is drained and the price-increase comparisons harden, revenue growth follows billings down toward the low teens.

Bull wins the mix; bear wins the rate. The honest read is a ~12–13% underlying business wearing a 16%-and-rising costume for a few more quarters. First round to the bear on growth sustainability, with the quality point banked for the bull.

Round 2 — Is the machine durable?

The bull's case here is genuinely strong. Autodesk owns the software that models buildings and manufactured parts; it is mission-critical, painful to rip out, and priced at 91% gross margins and a 41% non-GAAP operating margin. Growth is broad: Architecture/Engineering/Construction revenue up 17%, the newer "Make" (operations-leaning) line up 26%. Management points to construction strength, an enterprise digital-twin win with a major retailer, and an AI strategy built around connecting design data through construction and into decades of operations.

The bear concedes the moat and attacks the meter feeding it. A meaningful share of the growth is still price, and price is a machine with a ceiling — Autodesk has compounded list-price increases well above inflation for years, and each hike walks into a bigger base while renewals reminders grow louder. Construction is cyclical and interest-rate sensitive. And the earnings multiple is being paid on a number that isn't cash: GAAP EPS guidance of $7.89–$8.72 sits about $4 below non-GAAP's $12.52–$12.60, a standing bill for stock-based compensation that shows up as dilution rather than expense. The real earnings power is the $2.75 billion of free cash flow, and at roughly $13 a share that is 21 times — a quality price, not a cheap one.

Bull wins the economics; bear deducts for the SBC. No one reasonable fights the moat. The fight is what the moat is worth when it keeps compounding, and that is the valuation round.

Round 3 — The $3.6 billion it just spent

The bull sees MaintainX as the reason the growth story has a second act. Operations and maintenance consume the bulk of an asset's lifetime cost — management's framing is roughly 80% of lifecycle spend on the operate side versus 20% on design-build — and the operations software market is many multiples larger than Autodesk's core construction turf. MaintainX, a cloud maintenance-management provider with over $135 million in annual revenue growing above 50%, extends the "digital thread" into daily operations and feeds years of real-world data back into Autodesk's AI models.

The bear sees the price tag and the timing. ~$3.6 billion, roughly 30 times trailing revenue, for a business barely 2% of Autodesk's size, in a crowded market, paid partly with debt at a moment when the company's own guidance flags financing costs. The market's first verdict was cold: the stock fell about 4% on announcement day and slid to a 52-week low within two weeks. Buying growth at 30 times when you trade at 21 is an expensive way to admit the core engine needs a new lane.

Bear wins the price; the size saves the bull. At under 2% of revenue it cannot move the base case alone — but it tells you something about where the growth engine stands that the price likely should not.

What the price already demands

Put both stories through the same frame at $270.58. Roughly 21 times this year's guided non-GAAP EPS, about 21 times guided free cash flow, about 6.9 times sales. For that multiple to keep compounding, you need mid-teens growth to hold for two or three years and margins not to slip. If the billings and backlog numbers are the truth — and they have been the better signal through this whole transition — growth lands in the low teens, and 21 times is a fair-to-full price that cannot rerate much. Peer context does not rescue it: Autodesk trades at roughly a third the trailing multiple of Cadence (about 69 times) or Synopsys (about 83 times), but about a third above Adobe (roughly 16 times) or PTC (roughly 14 times). The premium over Adobe and PTC prices in the higher growth and fatter margin; the discount to the design-automation names prices in a thinner AI story. Mid-pack and defensible, but the market is not offering a margin of safety.

And the market has already collected part of the bet itself. The stock is up about 46% from its mid-June 52-week low, and the 6% pop on this report paid for confirming the guide raise. The bull's version — cleaner backlog, durable mid-teens — is now what is embedded in the tape, not what you get extra for believing early.

Ruling

The business case goes to the bull; the stock call at $270.58 goes to the bear, though not by a knockout. The reason is direction: the forward indicators are decelerating in a straight line — current RPO growth ran +23% at the end of fiscal 2026, +18% in the first quarter, +12% in the second; billings ran +30% for all of fiscal 2026 and +10% last quarter. Those are the numbers that feed revenue a year out, and they point down. Roll that forward and the listed ~21-times multiple is generous to a growth rate that is about to step down — while the market spent this quarter paying up for the beat.

The burden of proof is on the bull, and the tripwire is dated. At the fiscal Q3 report in late November: the ruling flips decisively back to the bull if billings re-accelerate to the mid-teens and current RPO growth holds around 14% or better — that would prove the backlog drawdown was a one-time cleanup ending, not a normalization beginning. The bear's call is confirmed if current RPO growth slides under ~10%, or if the non-GAAP operating margin slips below its ~39% full-year target as MaintainX costs cycle in. The backlog will tell you which story is true before the stock does.