Adobe closed its third fiscal quarter on August 28 and reported it after the bell on September 10: $6.76B in revenue, up 13%. The revenue line is not what the company led with. The two bullets sitting above it in the release are an AI growth rate and a headcount: one billion monthly active users across creativity and productivity.
Adobe put a user count where the money usually goes. That is the disclosure.
Two lines, two speeds
Subscription revenue for the quarter was $6.56B, up 14%. Adobe now reports it against two customer groups, and they are moving at different speeds:
- Creative & Marketing Professionals: $4.65B, up 13%
- Business Professionals & Consumers: $1.91B, up 16%
The professional line is still roughly 71% of subscription revenue. It is also the slower of the two. Total ARR exiting the quarter was $27.50B, and Adobe held its full-year target at 10.2% ending ARR growth.
The investor materials released alongside the filing carry the numbers that matter more. AI-first ending ARR passed $650M, up more than 150% year over year. Firefly ending ARR grew 40% quarter over quarter. Creative freemium monthly actives crossed 100 million, up more than 70% in a year.
Stack those against the ARR base and the quarter takes its actual shape.
$650M of AI-first ARR against $27.50B of total ARR is 2.4%.
The growth story Adobe put at the top of the release is currently paying for less than a fortieth of the company’s recurring revenue. In the same materials, the free user base added something on the order of 40 million people in twelve months.
Interim CFO Steve Day named the strategy without dressing it: expand the user base through freemium, deepen engagement with agentic experiences, convert that into durable growth. Nothing is hidden in that sentence. It just leaves one question open. Whose customers were those 40 million new free users?
The money has not arrived. The people have.
Three companies inside two weeks have moved the same variable in three different directions, which makes the comparison unusually clean.
Oracle reported Wednesday: $19.3B in revenue against a payroll down to 141,000, with almost all of the growth in rented racks and GPU hours and only 10% in the applications business that staffs consultants and support. That is a company shifting growth from the labor-intensive half of itself to the capital-intensive half.
Salesforce, at Goldman Sachs on September 9, moved the billing unit instead: a $550 bundle against a $1,300 stack, pricing off the seat.
Adobe did a third thing. No mass layoff, no repricing of the professional tier. It raised the floor. A hundred million people can now do masking, colour work, cuts and asset generation for free. Those are the tasks that used to bill by the hour or by the deliverable.
Raising the floor does not make the news. There is no WARN filing, no internal memo, no citable number of people let go. But for creative work the transmission path is more direct than a layoff, because the thickest layer of a studio’s revenue was never the creative judgment. It was volume: twelve crops of the same hero image, six cuts of the same spot, the weekly batch of social assets. That layer prices on one assumption: the client cannot do it themselves. Forty million new people learned to do it themselves this year.
Adobe’s own two lines are quantifying it. Professional buyers up 13%, everyone-else buyers up 16%, the free tier up 70%. Three numbers in a gradient, and the gradient slopes down.
What gets replaced, and when
The exposed unit is not the job title. It is the portion of design work that can be written down as a specification. The test is practical: if a task can be handed to someone else as one complete instruction — crop this hero into a nine-up, knock the background to pure white, export at three resolutions — it can already be written as one prompt. Three months ago that required a junior seat. Now the client’s own marketing associate ships it.
Ordered by that test, the exposure runs: e-commerce retouching and bulk export, multi-size social adaptation, template-driven short-form editing, stock-asset reworking. One layer up holds longer: layout judgment, brand consistency, the back-and-forth with a client who does not yet know what they want. Its input is context, not instruction.
The timing will not look like a layoff. It shows up first on quotes. Same job, $600 last year; this year the client holds two bids at $240, one of them from someone working a free tool. On the hiring side it appears as a quiet contraction in the number of junior openings, not the removal of existing ones. That contraction lands in no WARN filing and no Challenger monthly count, which means the data will see it about six months late.
We covered Artlist cutting 40% of staff to go AI-native at record revenue in June and Patreon cutting 20% while denying AI was the reason in July. Those were supply-side moves with countable headcounts inside a single company. Adobe’s quarter is the demand-side reading of the same shift, and it has no headcount in it at all. It has 100 million free monthly actives.
For an individual the usable move is one sort. Split current work into what can be written as one instruction and what cannot. The first pile keeps losing price over the next twelve months; do not deepen skill investment there. The second pile is what still prices: judgment, tradeoffs, alignment with a person.
The succession is the other half of the same disclosure
The same day, Adobe’s board named Anil Chakravarthy president and CEO effective December 1, with Shantanu Narayen moving to executive chair. Chakravarthy has spent the last six and a half years running Customer Experience Orchestration and worldwide field operations. He did not come up through the creative products.
That pairing is consistent with the numbers above: the successor comes from the enterprise and sales side of the house. The market’s reaction on the day was a roughly 2% after-hours decline, driven mostly by Q4 revenue guidance of $6.80B to $6.85B landing under expectations.
