Q3 Delivers the Q2 Wish List: AI-Influenced ARR Quantified at $5B+, DM ARR Growth Raised, Net New ARR Turning Point

Results vs. Consensus

Adobe's Q3 FY2025 print was a clean beat on every headline metric with stronger magnitude than Q2 — $80M revenue beat (+1.4% vs. Q2's +1.2%), $0.14 EPS beat (vs. Q2's $0.09), and $140M Digital Media ARR beat with +20bp acceleration in ARR growth (vs. Q2's +10bp). Beyond the headline, the disclosure framework upgrade is the more important development: AI-influenced ARR was quantified at $5B+ for the first time (the metric that frustrated the Street in Q2), and Digital Media ARR growth was raised from 11.0% to 11.3% (the guide reaffirmation that frustrated the Street in Q2). Both Q2 disappointments became Q3 catalysts.

Year-over-Year View

Sequential (QoQ) View

Revenue Assessment

The $80M revenue beat is the highest-magnitude beat against Street since Q4 FY24. The composition is constructive: Digital Media +11% YoY (+12% CC) accelerated modestly from Q2's +11% as-reported pace, with the +20bp acceleration in DM ARR growth (11.5% Q2 → 11.7% Q3 in CC) showing through. Business Professionals & Consumers subscription held at +15% YoY; Creative & Marketing Professionals subscription was +11% YoY (+10% CC), modestly above Q2's +10% (+11% CC). RPO acceleration from +10% YoY to +13% YoY is the most-overlooked positive metric — RPO is the forward indicator for subscription revenue, and a 300bp acceleration in RPO growth is meaningfully larger than the 30bp acceleration in current ARR growth. This implies the subscription growth in 2026 will outpace 2025.

Margin Assessment

Non-GAAP operating margin 46.5% held +50bp above guide (~46.0% midpoint), continuing the multi-quarter pattern of operating leverage despite the AI investment cycle. CFO Durn's explanation on the call — GPU fleet utilization, reserved-vs-on-demand instance mix optimization, algorithmic tuning of inference costs, and AI-driven internal employee productivity — is the most detailed margin commentary Adobe has provided on AI cost management. The Q4 operating margin guide of ~45.5% implies a ~100bp sequential step-down, which is investment-driven (Adobe MAX in October, year-end go-to-market push for CC Pro renewals, additional Firefly model integrations). The full-year operating margin will land ~46% — at the upper end of management's prior "mid-40s" framework — validating that AI cost pressure is manageable at the gross margin and operating leverage line.

EPS Assessment

Non-GAAP EPS $5.31 vs. Street $5.17 (+$0.14) — of which roughly $0.08 is operating beat, $0.04 is tax/share count (tax rate 18.5% per guide; ~$2.5B Q3 ASR reducing share count incrementally), and $0.02 is below-the-line. The raised FY25 non-GAAP EPS range of $20.80-$20.85 implies Q4 EPS of $5.35-$5.40 — directly aligned with the Q4 guide. The math is internally consistent: Adobe is no longer back-end-loading EPS into Q4 as aggressively as the prior Q2 guide implied. FY26 EPS sets up around $22.75-$23.50 depending on CC Pro mix uplift on renewals and AI-influenced ARR scaling.

Segment Performance

Both segments accelerated modestly on a constant-currency basis vs. Q2, with the bigger story being the customer-group acceleration within Digital Media. Adobe's segment reporting has been gradually augmented by customer-group disclosure (introduced at March 2025 FA Meeting, building scale this quarter) and is now further augmented by the GenStudio integrated-solution ARR disclosure (first time this quarter).

Customer Group Subscription Mix

Digital Media — $4.46B (+11% YoY)

Digital Media delivered a 20bp acceleration in ARR growth (11.5% Q2 → 11.7% Q3 in CC) and a 100bp acceleration in Creative & Marketing Professionals subscription growth (10% Q2 → 11% Q3 as reported). Net new DM ARR of approximately $500M sequential was the highest absolute net new ARR build in CY25 and — critically — the first quarter where net new ARR was flat-to-growing on a tougher YoY comp (a point one analyst explicitly framed as a "turning point quarter"). The drivers are well-distributed: Acrobat AI Assistant ending units +40% QoQ; AI Assistant conversations + summarizations +50% QoQ; Firefly App MAU +30% QoQ; first-time Adobe subscribers via Firefly app +20% QoQ. The Acrobat Studio launch (PDF Spaces + AI Assistant + Express integration at a premium tier) gives Adobe a new SKU above standard Acrobat AI Assistant, materially expanding the Business Pro & Consumer monetization stack.

"We exited the quarter with Digital Media ARR growing 11.7% year-over-year in constant currency driven by strong demand for higher-value AI-infused offerings including Creative Cloud Pro and Acrobat as well as AI-first products, including Firefly and Acrobat AI Assistant." — Dan Durn, EVP & CFO

Assessment: The Digital Media re-acceleration thesis is now operationally visible at the headline ARR level, not just the disaggregated sub-segment level. The 30bp FY25 DM ARR growth raise (11.0% → 11.3%) reflects management's confidence the acceleration sustains through Q4. CC Pro migration is "exceedingly well" per Narayen, and the renewal-cycle gating mechanism (most CC All Apps subs on annual cycles) means the ARR uplift compounds into FY26. If the Q3 pace of $500M sequential net new continues at this magnitude, exiting Q4 DM ARR lands around $19.1B (+13% YoY) — meaningfully above the 11.3% guide.

Digital Experience — $1.48B (+9% YoY)

DX subscription accelerated to +11% YoY, with the underlying AEP + GenStudio stack continuing to compound. AEP and native apps subscription growth held above +40% YoY (sustained from Q2). The first-time disclosed GenStudio integrated solution (Workfront + Frame + AEM Assets + Firefly Services + GenStudio for Performance Marketing) at $1B+ ARR growing >25% YoY is the most material new enterprise data point. Cross-cloud "One Adobe" deals grew +60% YoY in Q3, and 40% of Adobe's top 50 enterprise accounts have doubled ARR spend since the start of FY23 — an enterprise-expansion math validation that's well above what comparable enterprise software peers report. Adobe LLM Optimizer (early access, GA later in Q4) addresses the new category of LLM-based brand discovery, with Adobe's own Digital Index showing LLM traffic +4,700% YoY in July 2025.

"Our Workfront, Frame, AEM Assets, Firefly Services and GenStudio for performance marketing products, which are key components of the integrated GenStudio solution now exceed $1 billion in ARR growing over 25% year-over-year." — Shantanu Narayen, Chair & CEO

Assessment: The GenStudio $1B+ disclosure is the strongest enterprise-AI scale data point Adobe has provided to date. At $1B+ growing >25%, the integrated solution will reach $1.5-1.6B by exit FY26 — comparable in scale to mid-tier enterprise software franchises. The cross-cloud +60% YoY and top-50 doubled ARR data points reinforce that Adobe's enterprise expansion strategy is working: customers who buy into one Adobe cloud are increasingly buying into multiple, validating the integrated platform thesis. DX subscription is set up to outperform the guide by exit FY25.

AI-Influenced ARR — $5B+ (NEW)

For the first time, Adobe explicitly quantified AI-influenced ARR — revenue from products incorporating meaningful AI features — at $5B+, up from $3.5B+ exiting FY24. The implied YoY growth is approximately +43%, materially faster than total ARR growth (+11.7%). Adobe defines AI-influenced ARR as including CC flagship apps with AI features (Generative Fill in Photoshop, Generative Extend in Premiere Pro, Harmonize in Photoshop, Project Turntable in Illustrator), Acrobat AI Assistant subscriptions, Express premium tiers with AI features, and DX products with AI Assistant capabilities (AEP, AEM, GenStudio).

"Adobe AI influenced ARR surpassed $5 billion and we expect it to continue to rise as a percent of our business. Notably, ARR from our new AI-first products including Firefly, Acrobat AI Assistant and GenStudio for performance marketing has already achieved our end-of-year target of over $250 million." — Dan Durn, EVP & CFO

Assessment: The AI-influenced ARR quantification is the single most consequential disclosure of the quarter. It directly answers the Q2 frustration — analysts now have a benchmark to underwrite. At $5B exiting Q3 growing +43% YoY, the metric implies exit-FY25 AI-influenced ARR of ~$5.5-6.0B (~30% of DM ARR) and exit-FY26 of ~$7-8B (~38% of DM ARR by then). This is fast enough to bend the trajectory of the overall business meaningfully, and the +43% YoY growth in the AI-influenced bucket vs. +11.7% in total DM ARR mathematically requires non-AI ARR to be growing slower than total ARR — implying AI is the primary driver of marginal growth. The bears' "AI direct is only $250M, irrelevant" critique now sits next to the more important number: AI-influenced is 20x that, and accelerating.

Key Topics & Management Commentary

Overall Management Tone: Management was the most confident it has sounded in 12+ months, framing AI as the biggest opportunity for Adobe in decades and pointing to operational results (AI-influenced ARR quantification, GenStudio $1B disclosure, DM ARR raise) rather than aspirational framework. The tone shift from Q2's measured defense to Q3's assertive confidence was material and not subtle. Pushback was concentrated on second-order questions (margin sustainability, seat-vs-consumption shift) rather than primary AI-disruption concerns, signaling the AI overhang has lifted at least temporarily.

1. AI-Influenced ARR Quantification at $5B+ — The Q2 Frustration Fix

The single most consequential disclosure of the quarter. Adobe quantified what it had been calling "in the billions" qualitatively in Q2: AI-influenced ARR exceeded $5 billion in Q3, up from $3.5 billion exiting FY24. The +43% YoY growth in AI-influenced ARR vs. +11.7% in total DM ARR is the math that resets the AI narrative.

"Adobe is the leader in the AI creative applications category. Our AI influenced ARR has now surpassed $5 billion, up from over $3.5 billion exiting fiscal year 2024 and we've already surpassed our full year AI-first ending ARR target." — Shantanu Narayen, Chair & CEO

Assessment: The decision to quantify was directly responsive to the Q2 disclosure gap. At $5B+ growing +43% YoY, the metric is now testable: investors can track quarterly progress, model exit-FY26 magnitude, and benchmark against alternatives. The math implies AI features are now contributing ~$1.5B+ of net new ARR growth per year — significantly larger than the total +$2.0B annual DM ARR growth at the current trajectory — which means AI is the primary marginal growth contributor, not just an accelerant. This single number changes the bear/bull balance more than any other disclosure of the print.

2. Digital Media ARR Growth Raised from 11.0% to 11.3%

The second Q2-frustration fix: the FY25 DM ARR growth guide was raised from 11.0% to 11.3%. Adobe rarely raises ARR growth guides intra-year — the prior pattern is to reaffirm and over-deliver — so a 30bp raise mid-cycle is materially more bullish than the headline number suggests.

"Based on our momentum, we are pleased to raise our [Digital] media ending ARR growth target." — Dan Durn, EVP & CFO

Assessment: The 30bp raise plus the explicit explanation ("strong demand for higher-value AI-infused offerings including Creative Cloud Pro and Acrobat as well as AI-first products") puts a concrete data point under the AI-driven re-acceleration thesis. The raise also creates upside room — Adobe's historical pattern is to over-deliver on raised guides, which means exit-FY25 DM ARR growth could land at 11.5-11.7% in CC vs. the 11.3% raise. That math sets up an FY26 entry trajectory of 11.5%+ as the baseline.

3. The "Net New DM ARR Turning Point"

An analyst-led framing on the call captured Q3 as the first quarter of CY25 where net new DM ARR was flat-to-growing on a tougher YoY comp — the operational definition of an inflection. Sequential net new ARR of ~$500M in Q3 (vs. ~$460M Q2, ~$430M Q1 CC) is the data point validating the framing.

"It seems like this was kind of almost, I would say, a turning point quarter for you guys. It was the first quarter, I think, when net new DM ARR actually was flat to maybe even growing for the first time this year on a tougher comp."
— Alex Zukin, Wolfe Research (analyst framing during Q&A)

Assessment: The "turning point" framing — analyst-introduced, management-affirmed — is a clean inflection signal. Net new ARR growth on a tougher comp is the highest-quality form of acceleration evidence because comp pressure should be hardest precisely when AI-disruption fears are most pressing. The fact that Q3 reversed the multi-quarter trend of decelerating net new ARR growth is what gives Outperform-rating conviction beyond the disclosure-quality story alone.

4. GenStudio Integrated Solution Disclosed at $1B+ ARR Growing >25%

The first-time disclosure of GenStudio integrated solution ARR at $1B+ — combining Workfront, Frame.io, AEM Assets, Firefly Services, and GenStudio for Performance Marketing — anchors the enterprise content-supply-chain thesis with a concrete scale benchmark. At >25% YoY growth, the integrated solution is on track to reach $1.6-1.7B exit-FY26.

"Our Workfront, Frame, AEM Assets, Firefly Services and GenStudio for performance marketing products, which are key components of the integrated GenStudio solution now exceed $1 billion in ARR growing over 25% year-over-year." — Shantanu Narayen, Chair & CEO

Assessment: The GenStudio disclosure is the most-consequential enterprise data point Adobe has shared in 2025. At $1B+ growing >25%, this is a mid-tier enterprise software franchise within Adobe — scale comparable to early-stage public enterprise AI companies, but inside a much larger business. The pace of disclosure increase (Q2: "GenStudio +25% YoY" → Q3: "$1B+ >25% YoY") suggests management is becoming more confident in the trajectory and is willing to commit to benchmarks the market can underwrite. This is the right disclosure cadence for an inflection product.

5. Net New Customer Acquisition Through Firefly App

Firefly app continues to scale as the on-ramp to Adobe — first-time Adobe subscribers via Firefly grew +20% QoQ in Q3, with millions of mobile downloads since launch. Firefly App MAU grew +30% QoQ. The third-party model integration strategy is widening: Google Nano Banana (Gemini Flash 2.5) integrated day-of release; OpenAI GPT-image, Black Forest Labs Flux, Runway, Ideogram, Pika, Ray2, Luma all live.

"Firefly App MAU grew 30% quarter-over-quarter. Firefly App continues to attract next-gen creators with first-time Adobe subscribers through the app growing 20% quarter-over-quarter." — David Wadhwani, President, Digital Media

Assessment: The Firefly app as a top-of-funnel acquisition engine is working. The day-of integration of Google's Nano Banana is operationally important — it signals Adobe's positioning as the model-agnostic destination is fast enough to onboard new models at launch, removing the latency disadvantage that bears feared would let competitors race ahead. The +20% QoQ first-time-subscriber growth implies Firefly is bringing in roughly 0.5-1M new Adobe customers per quarter, a meaningful incremental acquisition channel.

6. Acrobat Studio Launch — New Premium Tier

Acrobat Studio launched in Q3 as a new premium tier combining Acrobat + AI Assistant + Express + PDF Spaces. PDF Spaces (new) transforms PDF collections into dynamic, conversational knowledge hubs — users can chat with multi-document collections to derive insights and remix content into emails, presentations, or other formats. Early reception was characterized as "strong" with "encouraging adoption and usage trends."

"The new Acrobat Studio includes PDF spaces which transforms collections of PDFs, web pages and other files into dynamic knowledge hubs that help people work smarter and faster using AI assistant to derive insights." — Shantanu Narayen, Chair & CEO

Assessment: Acrobat Studio is the highest-leverage SKU expansion Adobe has made in the Business Pro & Consumer category. It creates a premium price point above standalone Acrobat AI Assistant, capturing customers willing to pay for PDF Spaces + Express integration. If Acrobat Studio takes ~10% of net-new Acrobat subscribers at a 1.5-2x ARPU uplift, the contribution to Business Pro & Consumer subscription growth could be +200-300bp by exit-FY26. The +40% QoQ growth in Acrobat AI Assistant ending units in Q3 is the underlying conversion data point that gives confidence.

7. Adobe LLM Optimizer Launch + LLM Traffic Inflection

Adobe disclosed LLM-based traffic to U.S. retail sites grew +4,700% YoY in July 2025 (per Adobe Digital Index data on 1T+ visits to U.S. retail sites), and launched Adobe LLM Optimizer (early access in Q3, GA later in Q4) to help brands optimize content for LLM discovery. The product addresses the same pattern Anil Chakravarthy described from Adobe's own use of LLM Optimizer on adobe.com — customers increasingly ask product questions ("how do I edit this PDF") in ChatGPT/Perplexity rather than via search, and brands need to ensure their content surfaces in LLM responses.

"Adobe LLM Optimizer help shape how brands show up in [LLM] results which is drive influence, visibility and qualified traffic." — Anil Chakravarthy, President, Digital Experience

Assessment: LLM Optimizer is a category-creation product addressing a real, measurable, accelerating customer pain (+4,700% YoY LLM traffic growth). The pricing model and revenue ramp aren't disclosed, but the strategic logic is strong: AEM customers (Adobe's premium content management product) are the obvious deployment target, and LLM Optimizer creates a clear upsell SKU above standard AEM Sites. This is incremental TAM that didn't exist 12 months ago, and Adobe is positioned to capture it from a leadership position.

8. Margin Sustainability in the AI Investment Cycle

Operating margin held at 46.5% in Q3, +50bp above guide and well within the mid-40s framework. CFO Durn provided the most-detailed operating-cost commentary on AI to date — GPU fleet utilization, reserved vs. on-demand instance mix optimization, algorithmic inference cost tuning, and AI-driven internal employee productivity all cited as levers managing the cost curve.

"There's 2 vectors of productivity that the company is driving to underpin margin delivery. First one, how we drive GPU training fleets to support training, the utilization, the algorithms we use to efficiently get at model construction as well as continually loading that GPU fleet to make sure there's high utilization over time. The second piece is inferencing. Constantly tuning the algorithms and cost per inference. We watch this maniacally..." — Dan Durn, EVP & CFO

Assessment: The granularity of Durn's margin response — explicit references to GPU utilization and reserved vs on-demand instance pricing — is the strongest signal that Adobe is actively managing the AI infrastructure cost curve, not just absorbing it. This addresses one of the most-cited bear concerns (AI compute costs compress software margins toward 35-40%) with operational specificity. The 46.5% Q3 margin held against the +43% YoY AI-influenced ARR growth — which means the cost-of-AI-revenue is not currently scaling faster than AI-revenue, the fundamental margin sustainability condition.

9. Seats vs. Consumption Pricing Architecture

A nuanced question on whether AI shifts software pricing from seats to consumption — and whether that's a net headwind or tailwind for Adobe — drew a thoughtful response from Narayen positioning Adobe as winning in both scenarios: seat-based revenue continues to expand in the enterprise (Creative Cloud Pro renewals on existing seats); consumption-based revenue accelerates through Firefly Services + GenStudio (per-generation, per-asset monetization).

"We definitely view this as both as seat expansion as well as a marketing automation. And that's part of the reason why we — this customer grouping that we talk about, which is Creative Professional and Marketing Professionals... in the enterprise that's playing out exactly the way it is. It is actually still continuing to play out with seat expansion in the enterprise. But in addition to the seat expansion, the automation is definitely seeing a more rapid adoption." — Shantanu Narayen, Chair & CEO

Assessment: The "both/and" framing on seats vs. consumption is correct but not necessarily defensible long-term — if AI fundamentally automates creative work, seats SHOULD compress over time. The honest interpretation is that Adobe currently has both vectors growing because enterprise customers are buying more seats AND more consumption simultaneously (the +60% YoY One Adobe cross-cloud deals data point validates this). The longer-term question is whether the seat-expansion vector inflects negative at some point as AI reduces the number of humans needed per creative output — that's a 3-5 year question, not a 12-month one. For modeling purposes, both vectors growing is the right base case through FY27.

10. Retention Trends Improving with AI Adoption

A direct question on retention surfaced a clear pattern: AI feature usage correlates positively with retention. Customers using Firefly more, Acrobat AI Assistant more, and AI features in Creative Cloud more are renewing at higher rates and expanding usage faster. Commercial-safety + custom-model + workflow-integration combination is the moat that AI-native upstarts cannot easily replicate.

"This is the thing that we have seen is a direct correlation between increased use of AI and retention, and we feel very good about that. That's why we are so excited when we see the number of generations continuing to increase." — David Wadhwani, President, Digital Media

Assessment: The AI-usage-correlates-with-retention point is what bridges the AI-influenced ARR disclosure ($5B+) to the durability of Adobe's customer base. If AI features deepen retention, then the AI-influenced ARR is not just a leading indicator of growth — it's also a leading indicator of churn resistance. This is the most-overlooked positive of the call: AI investment is simultaneously a growth driver AND a defensive moat enhancer, not a choice between the two.

Guidance & Outlook

This is the second consecutive quarter Adobe has raised the FY25 guide. More importantly, the DM ARR growth raise — the metric the Street had pre-set for an upside surprise in Q2 and didn't receive — confirms management's confidence that the AI-driven re-acceleration is durable into Q4 and beyond. The Q4 revenue guide of $6.075-$6.125B implies sequential +2% QoQ, in-line with Adobe's normal Q4 seasonality.

Implied Q-over-Q ramp: Q4 revenue $6.10B midpoint = +1.8% QoQ from $5.99B. Q4 EPS $5.38 midpoint = +1.3% QoQ from $5.31. Both within normal Adobe seasonality.

Street at: Pre-print FY25 EPS consensus ~$20.50; post-print settles ~$20.85-$20.95 (modestly above midpoint).

Guidance style: The pattern has reverted to Adobe's classic "raise modestly each quarter, beat the raise next quarter." Two consecutive raises after a multi-quarter de-rating is the strongest forward signal management can provide short of issuing FY26 guidance.

Analyst Q&A Highlights

Adobe's Value-Add When Third-Party Models Dominate the Conversation

The opening line of questioning probed whether Adobe's role in the AI value chain becomes commoditized as third-party models (Google Nano Banana, OpenAI GPT-image, Black Forest Labs Flux) increasingly drive the "magic" in user-visible generations. Management framed Adobe's value as workflow integration and precision control — diffusion engines provide the generative capability; Adobe's applications provide the surrounding precision editing, composition, and production workflow.

Q: "Just on the demo that you guys gave on that video at the beginning, really highlighting the Adobe magic with what you're doing with Nano Banana — being able to manipulate images like that. Can you help us understand what part of that is the diffusion engine coming from Nano Banana? What part of that is the Adobe infrastructure underneath that, that enables that fine-grain manipulation?"
— Keith Weiss, Morgan Stanley

A: "The magic is clearly in our applications because we can take all of the models that exist and integrate that within our interface. And that's a completely nontrivial task of what we have done to build. That was actually the rationale for building Firefly because we understand whether they're diffusion or transformer models better than I think anybody can in the Creative Application. So I wouldn't underestimate the amount of magic that we have to make it look as seamless as it has."
— Shantanu Narayen, Chair & CEO

Assessment: The "workflow + precision control is the moat" framing is internally consistent with the Firefly app strategic positioning (model-agnostic destination). The risk is that as third-party models incorporate more workflow features natively, Adobe's surrounding-precision moat narrows. But near-term, the Photoshop + Nano Banana use cases shown on the call are operationally compelling and validate the integration thesis. The honest read: Adobe's moat in workflow + precision is real for 3-5 years; the durability beyond that depends on whether models stay model-only or expand into workflow.

Composition of the AI-First $250M+ Overdelivery

A direct question on what drove the early exceeding of the $250M AI-first products ARR target probed which sub-product (Firefly app, Acrobat AI Assistant, GenStudio for Performance Marketing) was outperforming. Management was reluctant to single out specific products (Narayen: "I always worry when I give 1 or 2 products, then the other teams will feel that I don't love them") but characterized the strength as broad-based across all three components.

Q: "Really great to see that you're exceeding that target for the full year of $250 million ARR from AI first products. There's a lot in there, and it would be great to get some color from you as to which component of that is driving the upside? And where are you seeing the strength in those AI-first products?"
— Brad Sills, Bank of America

A: "Firefly services, GenStudio. The overarching GenStudio solution is definitely seeing a lot of great adoption, AI Assistant in Acrobat and everything that we're doing associated with AI Acrobat... I always worry, Brad, when I give 1 or 2 products, then the other teams will feel that I don't love them a lot here within the company. But really, across all of them, I mean, we've seen significant momentum and easily crossed the $250 million mark."
— Shantanu Narayen, Chair & CEO

Assessment: The reluctance to disaggregate is a tactical choice that preserves disclosure flexibility — management doesn't want one product's slowdown to overshadow the others' growth. But operationally, the "broad-based strength" framing is consistent with the cross-cloud +60% YoY data point: customers buying Firefly Services are also buying Acrobat AI Assistant and GenStudio, which is the bundle effect Adobe's integrated platform thesis depends on. The right read: AI-first ARR is being driven by enterprise customers buying multiple AI products at once, not by a single hero product.

Net New ARR Inflection and PxQ Mix Going Into FY26

A question explicitly framed Q3 as a "turning point quarter" — first quarter of CY25 with net new DM ARR flat-to-growing on a tougher comp — and asked management to bifurcate AI-first adoption from CC Pro pricing/mix uplift in driving the inflection. Management characterized the strength as broad-based across all three monetization vectors (P × Q + V from the prior quarter's framework).

Q: "It seems like this was kind of almost, I would say, a turning point quarter for you guys. It was the first quarter, I think, when net new DM ARR actually was flat to maybe even growing for the first time this year on a tougher comp. You raised the guide. You exceeded your AI-first spending ARR target a quarter early. Can you help bifurcate like how much of this is just better adoption of the AI-first products versus PxQ? And ultimately, does this give you confidence in continuing to be able to grow DM ARR at least double digits for the foreseeable future?"
— Alex Zukin, Wolfe Research

A: "We've continued to execute, and we're now also starting to see that translate into monetization much like we thought we would... Creative Professionals, as we talked about, we're seeing generations accelerate. And the fact that they're using AI so much is what drove the strong migration to Creative Cloud Pro, really validated that strategy of the value and associated pricing with that... So that mix gives us a lot of momentum going into Q4."
— Shantanu Narayen, Chair & CEO and David Wadhwani, President, Digital Media

Assessment: The "turning point" framing was analyst-introduced but management didn't push back on it — a tacit acknowledgment that Q3 is an operational inflection. Adobe declined to issue FY26 guidance (Narayen: "we're not providing targets for next year") but the directional confidence (DM ARR raise, AI-influenced quantification, GenStudio $1B disclosure) signals management's view of FY26 is more bullish than the market is currently pricing. The honest read: Adobe is set up to issue FY26 DM ARR growth guidance at or above 11.5%, which would itself be a material catalyst at the Q4 print.

Seats vs. Consumption — The Long-Term AI Monetization Question

A pointed question on whether AI represents a structural headwind to seats (because AI automates work, reducing the number of humans needed) or a tailwind to consumption (because AI generation scales beyond human-driven workflows) drew the most thoughtful long-arc response of the call. Management positioned Adobe as winning in both vectors simultaneously through the customer-group disclosure framework.

Q: "There's a thesis out there for software in general — that AI is the headwind to seats and the seats will need to shift to consumption, the issue is then can [you] capture more consumption revenue than seat. How do you think about the relationship between seats and consumption in the Creative Cloud? Do you see that there's any pressure on the seat side? Do you see consumption as being upside to that?"
— Mark Moerdler, Bernstein Research

A: "We definitely view this as both as seat expansion as well as a marketing automation... It is actually still continuing to play out with seat expansion in the enterprise. But in addition to the seat expansion, the automation is definitely seeing a more rapid adoption. And so I think we're the one company that wins in either case and the fact that we have this unified offering that enables us to say, here's how — if you're the studio team within an enterprise, here's how you create these campaigns but then the velocity and agility of scaling that in a company through automation solutions."
— Shantanu Narayen, Chair & CEO

Assessment: The "both/and" answer is operationally correct in the near-term (the +60% YoY One Adobe deals data point proves customers are buying both seat-based and consumption-based products in the same contract) but is structurally fragile if AI eliminates a meaningful portion of creative seats over a 5-year horizon. The honest interpretation: Adobe is hedged on the seats-vs-consumption question by having both monetization vectors active, but the long-term mix shift toward consumption is real and Adobe will need to demonstrate consumption-revenue scale (currently <$1B in Firefly Services + GenStudio Performance Marketing) becomes >$3B by FY28 to fully hedge the seat-attrition risk. This is the right strategic posture for the current moment.

Margin Sustainability Against AI Compute Costs

A direct question on why operating margins are holding (or modestly expanding) against the AI investment cycle — when the bear thesis predicts AI compute costs compress software margins toward 35-40% — drew the most operationally specific response of the call. Durn cited four cost-management levers: GPU fleet utilization, reserved-vs-on-demand instance mix, algorithmic tuning of inference cost, and AI-driven internal employee productivity gains.

Q: "Dan, on margin, you've been speaking to mid-40s margin profile, you're still operating a bit above that this quarter. It looks like gross margins are actually up a touch versus last year. Why aren't you seeing degradation from AI adoption, given some of the metrics you're providing?"
— Michael Turrin, Wells Fargo Securities

A: "There's 2 vectors of productivity that the company is driving to underpin margin delivery. First one, how we drive GPU training fleets to support training, the utilization, the algorithms we use to efficiently get at model construction as well as continually loading that GPU fleet to make sure there's high utilization over time. The second piece is inferencing. Constantly tuning the algorithms and cost per inference. We watch this maniacally..."
— Dan Durn, EVP & CFO

Assessment: This is the most-specific AI-cost-management commentary Adobe has provided. The four-lever framework (GPU utilization, instance mix, inference tuning, internal productivity) is operationally robust and largely follows the AI cost-optimization playbook that mature AI-infrastructure operators have demonstrated works. The margin holding at 46.5% in Q3 against +43% AI-influenced ARR growth is the empirical proof — if AI compute costs were compressing margins, the Q3 print would have shown it. The right interpretation: Adobe's gross margin is structurally above 88% and operating margin in mid-40s is sustainable through at least FY27.

Adobe Experience Platform Agent Orchestrator — The Path to Agentic Enterprise AI

A question on the AEP Agent Orchestrator (launched first phase in Q3) probed Adobe's positioning in the agentic-software transition. Management framed agentic as a categorical SaaS-model shift — from human-driven SaaS to AI-driven agent orchestration — where Adobe builds first-party agents in its domain (audience, journey, data insights, product support) while also working with third-party agent orchestrators (Microsoft, ServiceNow) for inter-platform integration.

Q: "Anil, it's great to see agents for AEP now generally available. Can you talk us through the vision of how these agents drive greater [value] for customers? And then ultimately, how Adobe captures some of that benefit?"
— Brad Zelnick, Deutsche Bank

A: "We see this as a massive transformation from the SaaS model to the agentic software model... When you think of things like, for example, coming up with a new audience or a new journey, typically, you needed to have somebody with some data expertise to do that. Somebody who could come up with a SQL query. So usually, marketing teams set up alternate teams or a dedicated team to do that kind of work. What we see now with Agent Orchestrator is that enables the creation of an audience agent, a data insights agent, a journey agent that now lets marketers with a conversational interface get access to that kind of data."
— Anil Chakravarthy, President, Digital Experience

Assessment: The agentic positioning is structurally important. Adobe's domain knowledge (decades of marketing, content, and customer-experience data infrastructure via AEP) gives it advantages in building purpose-built marketing agents that horizontal agent platforms (LLM-only chat interfaces) cannot easily replicate. The right read: AEP Agent Orchestrator is a defensive moat against horizontal LLM platforms (ChatGPT, Gemini) trying to enter marketing analytics, and it creates a new ARR layer on top of existing AEP subscriptions. The pricing model isn't disclosed but the strategic logic is sound.

Retention and Commercial Safety as the Long-Term Moat

The closing question linked AI usage to retention metrics and to the commercial-safety positioning. Wadhwani disclosed a direct positive correlation between AI feature usage and customer retention — the more AI features a customer uses, the more they renew and expand. The commercial-safety wrapper (IP-cleared Firefly Foundation Models) is increasingly cited by enterprise customers as the differentiating factor for production-grade AI generation.

Q: "I think a resounding theme on this call is the infusion of AI into the core applications is really what's differentiating. Can you talk about what you're seeing in terms of retention trends as AI adoption grows? And separately, we've also started to see some noise out there about intellectual property issues when it comes to AI usage. Maybe talk about how important the commercial safety aspect is as you spend time with enterprises."
— Saket Kalia, Barclays

A: "This is the thing that we have seen is a direct correlation between increased use of AI and retention, and we feel very good about that. That's why we are so excited when we see the number of generations continuing to increase... [Commercial safety] lets us go from that seat-based pricing to a value-based pricing. And what the underpinning of that is, is the ability to have these commercially safe models. On top of that, the ability to create customizations of those models for your specific needs."
— David Wadhwani, President, Digital Media

Assessment: The AI-usage-correlates-with-retention disclosure is the most-overlooked positive of the call. If AI features deepen retention, then the $5B+ AI-influenced ARR is not just leading-indicator-of-growth — it's also leading-indicator-of-churn-resistance. Combined with the commercial-safety positioning (where Adobe's IP-cleared foundation models are increasingly demanded by enterprise customers wary of copyright exposure), this creates a flywheel: AI usage → higher retention → larger installed base → more AI usage → wider moat. This is the strongest long-term moat data point Adobe has surfaced in CY25.

What They're NOT Saying

1. FY26 DM ARR Growth Guidance: Narayen explicitly declined to provide FY26 targets ("we're not providing targets for next year"). Given the FY25 raise to 11.3% DM ARR growth and the underlying trajectory (RPO +13% YoY, AI-influenced +43% YoY), the implied FY26 setup is ~11.5-12.5% DM ARR growth — but management is holding the explicit guide for the Q4 print. The non-disclosure preserves flexibility for a potential FY26 raise to be the Q4 catalyst.
2. CC Pro Renewal Mix Disclosure: No specific data point on what % of CC All Apps renewals are converting to CC Pro at the higher price tier. Wadhwani referenced "strong migration" and Narayen said CC Pro is the "sweet spot" — but no actual conversion rate. The disclosure cadence suggests Adobe is waiting to provide this number once the international rollout has more renewal data behind it.
3. AI-Influenced ARR Sub-Segments: The $5B+ AI-influenced ARR was disclosed as a single number without breaking down by product (CC flagship vs. Acrobat vs. DX). Specific sub-segmentation would let investors model which products are driving the +43% YoY growth — but Adobe is holding the granularity for now.
4. GenStudio Customer Concentration: The $1B+ GenStudio integrated solution ARR was disclosed without customer concentration data (e.g., top-10 customer % of total). Given the cross-cloud "One Adobe" +60% YoY data point, concentration is likely modest, but the absence of detail leaves the durability question open.
5. Express Standalone MAU Confirmation: Adobe again declined to confirm Express standalone MAU (the Q2 analyst-derived estimate was ~50M). The combined Acrobat + Express MAU at ~25% YoY growth was reaffirmed (with a minor discrepancy between Narayen's "20% YoY" and Durn's "25% YoY" — likely rounding, but unusual). The non-disclosure of Express MAU continues to preserve flexibility.
6. LLM Optimizer Pricing/Monetization Model: Adobe LLM Optimizer was launched in early access with no disclosure on pricing model (per-seat, per-asset, consumption-based?) or expected FY26 revenue contribution. The product is currently positioned as an AEM-attached SKU but the monetization architecture isn't yet visible.
7. Capital Return Pacing for Q4 and FY26: Q3 buyback was $2.5B vs. Q2's $3.5B — a modest deceleration. No commentary on whether Q4 buyback will return to the $3.5B level or settle in the $2-3B range. The $8.4B remaining authorization gives ~3-4 quarters of buyback at current pace, suggesting a new authorization is likely at the Q4 print or shortly after.
8. Adobe MAX (October) Catalyst Magnitude: Multiple references on the call to Adobe MAX in October as a showcase for new innovations across creativity, marketing, and AI. The framing suggests substantive new product launches but no preview of magnitude. The Q4 print (December) will include both Adobe MAX disclosures and FY26 guidance — a single-quarter catalyst stack that is unusual.

Market Reaction

• Pre-print setup: ADBE closed at ~$355 entering print; -20% YTD vs. S&P +6%; -30% trailing 12 months; trailing 30-day +5% (modest recovery from Q2 sell-off); options-implied ±6-8% move; trading at ~17x forward EPS (vs. ~30x 5Y average). Sentiment cautious-but-improving.
• After-hours move: Initial +6-8% on print headlines (AI-influenced ARR quantified at $5B+, DM ARR growth raised, AI-first $250M target exceeded a quarter early). Held the gains on call as the disclosure framework upgrade landed. Closed after-hours +7-8% near $383.
• Next-day session (September 12): Opened +6% near $376 and closed at approximately $385, **up +8.5% (+$30 from Sept 11 regular-session close).** Volume ~14M shares, roughly 2.2x trailing 30-day average. Largest single-day rally for ADBE in CY2025.
• Peer reaction: Software peers participated more modestly; Salesforce +1%, Snowflake +2%, Datadog flat. The rally was Adobe-specific, reflecting the AI-disruption narrative reset rather than a broader software-sector signal.

This is the print that resolves the multi-quarter AI overhang. The combination of (a) quantified AI-influenced ARR at $5B+ (the Q2 disclosure gap fix), (b) DM ARR growth raised from 11.0% to 11.3% (the Q2 reaffirmation frustration fix), (c) AI-first products exceeding the $250M target a quarter early, and (d) GenStudio integrated solution disclosed at $1B+ — addresses every major bear concern that has compressed the multiple from 30x to 17x over the past 18 months. The market's reaction is proportionate to the narrative reset, not just the operational beat.

What's also constructive is the second-derivative signals: the net new DM ARR inflection (first flat-to-growing quarter in CY25 on a tougher comp), RPO acceleration to +13% YoY (from +10%), and margin holding above mid-40s framework. These are durability signals, not just headline strength — they suggest the AI-driven re-acceleration thesis has gone from "possibly happening, hard to verify" to "verifiably happening, with measurable forward momentum."

At ~$385, ADBE trades at ~18x FY26 EPS. The path to ~$500-$520 by year-end FY26 requires the multiple to expand to ~22-23x as AI-influenced ARR continues to compound — a step that, given the $5B base growing +43%, no longer requires heroic assumptions. The bear case (multiple compresses further to mid-teens) requires AI-influenced ARR growth to decelerate sharply, which the underlying trajectory does not suggest. The risk/reward at $385 is asymmetric enough to support an Outperform call.

Street Perspective

Debate: Is AI-Influenced ARR the Right Metric, or Is It a Definitional Stretch?

Bull view: The $5B+ AI-influenced ARR is a real, defensible metric — Adobe defines it narrowly as ARR from products with meaningful AI features (Generative Fill in Photoshop, Acrobat AI Assistant, AEP AI Assistant), not all-of-Creative-Cloud. At +43% YoY growth (from $3.5B to $5B+) vs. +11.7% total DM ARR growth, AI-influenced is mathematically driving the marginal acceleration. This is the right metric for the AI-transition era and Adobe is showing leadership by disclosing it.

Bear view: "AI-influenced" is a definitional stretch — Adobe is including products that have AI features alongside many non-AI features (e.g., CC All Apps subscription is counted as AI-influenced even though most users may not be using AI features). The $5B number is therefore not directly comparable to peers' AI revenue disclosures, and the +43% YoY growth could partly reflect re-categorization rather than genuine AI-driven growth.

Our take: The definitional concern is valid in principle but small in practice. Even if half of "AI-influenced" ARR is generously counted, the underlying $2.5B+ at +30%+ YoY is still substantial. More importantly, the metric is consistent quarter-over-quarter (Adobe will report $5B-and-growing in Q4, then track it through FY26), so the trajectory is what matters, not the absolute level. The disclosure is directionally correct and the right framework for an AI-transition era. We side with the bull camp.

Debate: Does the Adobe MAX + Q4 Print + FY26 Guide Stack Justify a Pre-Print Buy?

Bull view: Adobe MAX in October will showcase substantive new AI products (likely including LLM Optimizer GA, additional Firefly model integrations, and new Acrobat Studio capabilities). The Q4 print in December will issue FY26 guidance — likely DM ARR growth of 11.5-12% based on Q3 trajectory — which would be a material positive catalyst. The stack of MAX + Q4 print + FY26 guide creates a multi-month tailwind environment, making pre-print positioning attractive.

Bear view: The October MAX event has historically been more PR than substance for the stock price — Adobe has held annual MAX events for years with little persistent stock impact. The FY26 guide could come in below expectations if management chooses to anchor conservatively after two consecutive intra-year raises (the "set low to beat consistently" pattern). The stock has already moved +8.5% on the print; chasing it pre-MAX risks giving back the gain if MAX disappoints.

Our take: The bear concern about MAX impact is empirically correct historically but the current AI moment may be different — MAX is the natural venue for Adobe to demonstrate that the AI capability stack is genuinely differentiated, and post-Q3-narrative-reset, the market may be more responsive than in prior years. The FY26 guide is the bigger catalyst — at 11.5%+ DM ARR growth, the stock has room to move toward $500; at 11% (flat with FY25), the stock would likely consolidate. We see net upside in the stack, with the larger move coming at the Q4 print, not at MAX itself.

Debate: Does the Multiple Re-Rate to 22-25x or Stay Compressed at 17-19x?

Bull view: The 17x forward EPS multiple was set in the AI-disruption-fear environment of 2024-Q2 2025. With AI-influenced ARR now quantified at $5B+ growing +43% YoY, the multiple should re-rate to reflect Adobe's positioning as an AI-winner-not-loser. Peer comparable AI-driven software franchises trade at 25-30x (Salesforce, ServiceNow, Snowflake post-AI). A re-rate to 22-23x on FY26 EPS of ~$23 puts ADBE at $500-$525, +30-35% from current.

Bear view: The 17x multiple reflects structural concerns about Adobe's growth durability beyond CY26 — even if AI re-accelerates ARR in the near-term, the long-term seats-vs-consumption question (whether AI compresses seats faster than consumption scales) remains unresolved. The mature-software-cohort de-rating that has compressed Atlassian (15x), Workday (18x), Salesforce (22x) is structural, not cyclical, and Adobe sits in that cohort.

Our take: The bull view has the better near-term tailwind argument — the AI-narrative reset is real and the multiple has room to expand 200-300bp over the next 2-3 quarters. The bear view has the better long-term concern — the seats-vs-consumption question is genuine and will weigh on the multiple over a 3-5 year horizon. We expect ADBE to trade in a 19-23x forward EPS range through FY26, with the multiple expansion likely concentrated in 1H26 as AI-influenced ARR continues to compound. PT range of $440-$500 fits this framework.

Model Update Needed

Valuation impact: Updated PT range: $460 base / $520 bull / $380 bear. Base case ($460) assumes 20x FY26 non-GAAP EPS of ~$23.00. Bull case ($520) assumes 22.5x on FY26 EPS of ~$23.30 with continued AI-influenced ARR compounding. Bear case ($380) assumes 16.5x on $23.00 with mature-software-cohort de-rating. At ~$385 post-print, base case implies +19% upside; bull +35%; bear -1%. Up/down ratio ~3:1 — Outperform.

Thesis Scorecard Post-Earnings

Overall: Thesis materially strengthened. Four bull points confirmed (one new — the retention correlation); three of four bear points mitigated or partially mitigated. The disclosure framework upgrade ($5B AI-influenced ARR + $1B GenStudio + 11.3% DM ARR raise) directly addresses the bear case's strongest critique (AI is too small to matter). The 220bp creative-subscription acceleration validated in Q2 continued and accelerated in Q3 (Creative & Marketing Pro subscription +11% Q3 vs. +10% Q2). The "net new DM ARR turning point" framing on the call captured an operational inflection that justifies an explicit upgrade.

Action: Upgrading to Outperform from Hold. The combination of (a) AI narrative reset via disclosure quantification, (b) confirmed re-acceleration in net new DM ARR, (c) margin sustainability proof, and (d) attractive 18x FY26 EPS multiple creates an asymmetric setup. Position sizing for portfolios already holding can scale up; new positions can initiate at current levels. Next catalysts: (1) Adobe MAX in October, (2) Q4 print in December with FY26 guidance, (3) CC Pro renewal data exiting FY26 H1.