From Aryan Vatsa | Product & Market Analysis
Productivity Software in 2026: What Five Vendors' Seat Disclosures Show
On this page
Microsoft 365 commercial paid seats grew 6% year over year in every quarter from mid 2025 to mid 2026, while the revenue on those seats grew more than twice as fast. That gap is the productivity software market in 2026 in one line. Across five vendors that disclose seats or subscribers in some form, agent features have not produced a seat boom. They have produced a pricing story, and in one case a genuine expansion story.
Key takeaways
- Microsoft 365 commercial seats grew 6% for five quarters running. Microsoft 365 commercial cloud revenue grew 17% for fiscal 2026, so price and mix carried most of the growth, not new seats.
- Atlassian is the only one of the five where seat growth visibly accelerated revenue. Cloud revenue growth rose from 25% to 31% across six quarters, and management credited paid seat expansion alongside price increases.
- Dropbox paying users barely moved in 18 months. The count ran from 18.16 million to 18.19 million, with three quarters of decline before three quarters of small gains.
- The AI seat is reported separately from the base seat. Microsoft 365 Copilot reached over 30 million paid seats by June 2026, against a commercial base Microsoft put at over 450 million in January.
The short answer. Did AI features grow productivity software seats between early 2025 and mid 2026? Mostly no. Microsoft held a steady 6% seat growth rate, Dropbox held flat, and Zoom grew its large customer cohort at about 8%. Atlassian grew seats faster. Revenue grew faster than seats at four of the five, which points to price, mix and AI add-ons.
Productivity software seats: six quarters from five vendors
Every number below comes from a company's own earnings release, shareholder letter, investor metrics page or periodic filing. Nothing is modelled. Where a cell says "not disclosed", the company did not publish that figure for that quarter, and that absence is itself part of the finding.
The five vendors measure five different things. Only Dropbox publishes an absolute count of paying users every quarter. Microsoft publishes a growth rate. Zoom and Atlassian publish counts of large customers. Alphabet publishes a consumer subscription total when it chooses to. Read across a row with care, and never add a column.
| Vendor and metric | Mar 25 | Jun 25 | Sep 25 | Dec 25 | Mar 26 | Jun 26 |
|---|---|---|---|---|---|---|
| Microsoft: M365 Commercial paid seat growth y/y [1] | 7% | 6% | 6% | 6% | 6% | 6% |
| Microsoft: M365 Consumer subscribers [2] | 87.7M | 89.0M | Not disclosed | Not disclosed | Not disclosed | Not disclosed |
| Microsoft: M365 Copilot paid seats [3] | Not disclosed | Not disclosed | Not disclosed | 15M | See note | Over 30M |
| Alphabet: paid subscriptions, consumer [4] | Over 270M | No total given | Over 300M | Over 325M | 350M | No total given |
| Zoom: customers over $100K TTM revenue [5] | 4,192 | 4,274 | 4,363 | 4,468 | 4,534 | 4,625 |
| Atlassian: customers over $10K Cloud ARR [6] | 50,715 | 51,978 | 53,017 | 55,369 | 55,913 | 57,334 |
| Dropbox: paying users [7] | 18.16M | 18.13M | 18.07M | 18.08M | 18.09M | 18.19M |
Definitions. [1] Microsoft counts paid users covered by a Microsoft 365 Commercial subscription at period end; it publishes growth, not a total, though it said "over 450 million" in January 2026. [2] Microsoft removed this metric after the June 2025 quarter. [3] Microsoft gave 15 million on its January 2026 call and "over 30 million" in its July 2026 release; a March 2026 figure appears in third-party coverage but we could not confirm it in Microsoft's own materials. [4] Alphabet counts paid subscriptions across consumer services, led by YouTube and Google One; "over" figures are lower bounds; it is not a Workspace seat count. [5] Zoom's quarters end one month later than the column heading (April, July, October, January). [6] Atlassian's fiscal year ends 30 June; columns are aligned to calendar quarter ends. [7] Dropbox counts users with active paid licences.
Three patterns jump out before any vendor detail. The large-customer cohorts at Zoom and Atlassian rose every single quarter. The broad-base counts at Microsoft and Dropbox did not accelerate at all. And two of the five vendors published less in mid 2026 than they did in early 2025.
Microsoft 365 seats held at 6% while revenue ran ahead
Microsoft is the benchmark because it has the largest paid seat base in productivity software. Its investor metrics define the seat measure as paid users covered by a Microsoft 365 Commercial subscription at the end of the period. The growth rate read 7% in the March 2025 quarter and 6% in each of the five quarters after it, per Microsoft's own metrics page.
Revenue tells a different story. Microsoft 365 commercial cloud revenue grew 17% for fiscal 2026 as a whole, against seat growth of 6%. In the June 2026 quarter the reported figure was 14%, which Microsoft's release said was 16% adjusted for a prior-year revenue recognition benefit.
Microsoft's own earnings remarks said the installed base grew primarily in small and medium business and frontline worker offerings. Those are typically lower-priced seats. So the seat mix was drifting toward cheaper plans while revenue per user still rose.
That only works if existing customers are paying more per seat. Microsoft's annual report attributes commercial revenue to installed base growth and average revenue per user expansion. The 2026 list price increase on commercial suites and upgrades to higher tiers both feed that second term. My reading: Microsoft's productivity growth in this period was a price and mix story, not AI-driven seat growth.
Copilot is a second seat, not more seats
Microsoft first gave a Microsoft 365 Copilot number on its January 2026 call: 15 million paid seats, as Computerworld reported. On the same call it cited over 450 million commercial seats. By the July 2026 release, Satya Nadella said Copilot had reached over 30 million paid seats.
Copilot is an add-on licence. Every Copilot seat sits on top of an existing Microsoft 365 seat, so it cannot raise the base seat count at all. It raises revenue per user. At 30 million against a base above 450 million, the attach rate is under 7%. That is real growth in a new revenue line and a small share of the installed base. The price-per-seat side of this is covered in the comparison of what the $30 Copilot seat buys against Gemini.
The consumer number that disappeared
Microsoft reported 87.7 million Microsoft 365 Consumer subscribers in March 2025 and 89.0 million in June 2025. From the September 2025 quarter onward, that row is gone from the metrics page. The fiscal 2026 annual report describes consumer subscriber growth as a percentage, not a total.
Companies remove metrics for many reasons, including simple relevance. The pattern still matters. Consumer revenue growth accelerated into the 20s during fiscal 2026 at the same time the subscriber count stopped being published, which means price is again doing much of the work.
Atlassian paid seats are the one clear expansion story
Atlassian stopped publishing an exact total customer count and now uses rounded wording such as "more than 350,000". The metric it does publish every quarter is the number of customers paying over $10,000 in Cloud annualised recurring revenue. That count went from 50,715 in March 2025 to 57,334 in June 2026, about 13% across five quarters.
The revenue line moved faster and, unusually, accelerated. Cloud revenue growth ran 25%, 26%, 26%, 26%, 29% and 31% across the six quarters, according to the back data in the Q4 fiscal 2026 shareholder letter.
What Atlassian says drove it
The December 2025 letter said paid seat expansion rates "remain stable" and named seat expansion within existing customers as the first driver of 26% cloud growth. The June 2026 letter credited "strong seat expansion in core Jira and Confluence". The fiscal 2026 annual report attributes subscription revenue growth to paid seat expansion from existing customers and price increases.
Note the order of that sentence. Seats first, price second. Atlassian is the only vendor in this set where the company's own language puts seat growth ahead of pricing, and where revenue growth accelerated rather than drifting.
There is an important asterisk. Atlassian announced end of life for its Data Center product in September 2025, which pushes self-managed customers onto Cloud. Some of the cloud seat growth is therefore migration of existing seats from one Atlassian product to another, not net new users. The company's guidance for fiscal 2027 explicitly expects muted seat expansion in Data Center as customers prepare to move.
The AI layer shows up as a multiplier, not a headcount. The December 2025 letter said Rovo passed 5 million monthly active users and that customers using AI code generation expand seats 5% faster than those that do not. That is the closest any of the five vendors comes to claiming AI produces more seats, and it is a self-reported correlation.
Zoom and Dropbox: seats held flat behind stable revenue
These two companies are the closest thing to a control group. Both sell productivity software with a large self-serve base, both have shipped AI features, and both have seen almost no movement in the base.
Zoom grows the top of the base and holds the rest
Zoom does not publish seat counts or a total enterprise customer count in its earnings releases. It defines a customer as a distinct buying entity, which can be one paid user or a whole organisation. What it publishes is the count of customers above $100,000 in trailing 12-month revenue, which rose from 4,192 to 4,625 over the six quarters, per Zoom's fiscal Q2 2027 release.
The rest of the picture is stable rather than growing. Online average monthly churn sat between 2.7% and 3.0% in every quarter. The trailing 12-month net dollar expansion rate for enterprise customers was 98% for four quarters, then 99% for two. A rate below 100% means the existing enterprise base, in aggregate, spent slightly less than a year earlier.
Zoom's AI metrics are growth rates without bases. The fiscal Q1 2027 release reported AI Companion paid users growing 184% year over year, with no absolute count. A percentage without a base cannot be weighed against a seat count, and I would not try.
Dropbox: three quarters down, three quarters up
Dropbox is the only vendor here that publishes an absolute paid user count every quarter. It fell from 18.16 million in March 2025 to 18.07 million in September 2025, then rose for three quarters to 18.19 million in June 2026. The June 2026 quarter added 96,000 paying users, which the company called its third consecutive quarter of growth.
Revenue per user did not pick up the slack. Average revenue per paying user moved between $138.32 and $141.18 across the period. The 10-Q attributes the year-over-year gain to favourable exchange rates and a shift to monthly plans, partly offset by more customers choosing lower-priced plans such as Simple. In plain terms, Dropbox's recovery in user count is coming from the cheap end.
Alphabet subscriptions grew, but they are not productivity seats
Alphabet is in this set because it discloses a subscription number and because Google Workspace is the main rival to Microsoft 365. The honest finding is that Alphabet discloses nothing about Workspace seats at all. In the filings and call remarks we reviewed, there is no Workspace customer count, no Workspace seat count and no Gemini Enterprise paid seat count.
What it does publish is a total of paid subscriptions across consumer services. That went from "surpassed 270 million" in April 2025 to over 300 million in October 2025, over 325 million in February 2026 and 350 million in April 2026, per Alphabet's Q1 2026 release. The Q2 2026 call gave no new total.
Google One AI plans are the productivity angle
The part of this that touches productivity software is Google One. Alphabet's CFO said on the July 2026 call that subscriptions, platforms and devices revenue grew 15% to $12.9 billion, with Google One growth "driven by demand for AI plans". That is consumer AI bundled into storage and productivity features, sold per person rather than per enterprise seat.
It is a real signal that people will pay for AI inside a productivity bundle. It is not evidence about enterprise seat counts, and treating 350 million subscriptions as comparable to Microsoft's 450 million commercial seats would be an error. Most of those subscriptions are YouTube.
The disclosure retreat is the quiet finding
Lining up six quarters exposes something a single earnings report hides. The metric set is moving, and mostly toward less precision on the base and more precision on the AI layer.
There are innocent reasons. A metric can stop being useful, or a business can be reorganised so the old line no longer maps to anything. But investors and buyers should apply one simple rule. When a company stops publishing a count and keeps publishing the revenue, assume the count was no longer flattering until shown otherwise.
I hold that view firmly for one reason. Seat counts are the one input that tells you whether software is reaching more people. Revenue can rise on price alone, and the autopsy of per-seat pricing shows how far price can carry a business whose user base is not growing.
| Vendor | Verdict | Evidence from the filings | Strongest counterpoint |
|---|---|---|---|
| Microsoft | Held seats, grew price | 6% seat growth flat for five quarters; 17% fiscal 2026 revenue growth | 6% on a base above 450 million is still tens of millions of seats |
| Atlassian | Grew seats | Cloud growth accelerated to 31%; company names seat expansion first | Data Center end of life moves existing seats into Cloud |
| Zoom | Held | Net dollar expansion 98% to 99%; churn steady near 2.9% | Large customer count grew every quarter |
| Dropbox | Held, cheaper mix | Paying users flat at about 18.1 million; growth from lower-priced plans | Three straight quarters of user growth after a decline |
| Alphabet | Not measurable for seats | Consumer subscriptions grew; no Workspace seat data | Google One AI plans show willingness to pay |
Where this analysis is weakest
This post lines up five different metrics and draws a pattern from them. That method has limits, and they should be stated plainly before anyone quotes the verdict column.
The metrics are not comparable
A Microsoft seat growth rate, a Zoom large-customer count and a Dropbox paying user count measure different things. A customer crossing $100,000 at Zoom can reflect price increases as easily as more seats. A customer crossing $10,000 at Atlassian can reflect a migration rather than an expansion. The table shows what each company chose to disclose, not a common measure of adoption.
Six quarters is short, and agents are new
Most agent products in this set reached general availability during or shortly before this window. Seat effects from a new feature can take several renewal cycles to show up, because enterprise contracts renew annually or longer. A flat seat line through mid 2026 does not prove agent features will never grow seats. It shows they had not done so yet in what these companies disclose.
The opposite risk also exists. If agents do work, they may reduce the number of humans who need a seat, which is the compression case argued in the analysis of seat compression and SaaS pricing. Flat seats could be the early stage of that, or simply a saturated market. These filings cannot tell the two apart.
Price is inferred, not disclosed
None of the five companies publishes a clean split of revenue growth into seats and price. The "price did the work" reading for Microsoft comes from subtracting a seat growth rate from a revenue growth rate, plus the company's general statement about revenue per user. It is an inference, sound in my view, not a disclosed fact.
How to read the next productivity software filing
The next round of results covers the September 2026 quarter. Here is what would change the conclusions above, vendor by vendor.
For Microsoft, a seat growth rate above 6% would be the first sign that Copilot or agent bundles are pulling in new seats rather than monetising existing ones. A drop below 6% while revenue keeps growing would strengthen the pricing reading. Watch whether the Copilot number is given again, and whether it is given as a count or a growth rate.
For Atlassian, the question is whether cloud growth holds once the Data Center migration wave passes. Guidance of about 25.5% cloud growth for fiscal 2027 already implies some deceleration. For Dropbox, a fourth straight quarter of paying user growth would matter more than any AI announcement. The framing in the piece on value metrics that survive automation is useful here: ask which metric a vendor would bill on if seats stopped growing.
For buyers, the practical point is simpler. Your vendor's revenue is growing faster than its seat base, which means your renewal price is the growth engine. Plan negotiations on that assumption, and read the analysis of hybrid pricing and the retention gap before your next renewal call.
Frequently asked questions
How many Microsoft 365 commercial seats are there in 2026?
Microsoft said in January 2026 that it had over 450 million paid Microsoft 365 commercial seats. It does not publish the total every quarter. Instead its investor metrics report seat growth, which was 6% year over year in every quarter from June 2025 through June 2026. Any later total you see is an extrapolation by a third party, not a Microsoft disclosure.
Is the productivity software market growing in 2026?
Revenue is growing, seats much less so. Microsoft 365 commercial cloud revenue grew 17% in fiscal 2026 against 6% seat growth. Atlassian cloud revenue grew 31% in the June 2026 quarter. Dropbox revenue grew under 1% with flat paying users. The market is growing mostly through higher prices, richer tiers and AI add-ons, with Atlassian the main exception showing strong seat expansion.
How many paid Microsoft 365 Copilot seats are there?
Microsoft reported over 30 million paid Microsoft 365 Copilot seats in its July 2026 earnings release, up from 15 million disclosed on its January 2026 call. Copilot is an add-on to an existing Microsoft 365 seat, so these are not new users. Against a commercial base Microsoft put above 450 million, the attach rate is under 7%.
Does Atlassian report paid seats?
Atlassian does not publish a paid seat count. It reports qualitative statements about paid seat expansion, which it called stable in December 2025 and strong in June 2026. Its quantitative customer metric is the number of customers with more than $10,000 in Cloud ARR, which reached 57,334 at 30 June 2026. It rounds its total customer count to more than 350,000.
Why did Microsoft stop reporting Microsoft 365 Consumer subscribers?
Microsoft has not explained the removal in its metrics materials. The count was 89.0 million at June 2025 and the row disappeared from the following quarter. The annual report now describes consumer subscriber growth as a percentage. Consumer revenue growth accelerated over the same period, which suggests price increases contributed more than subscriber growth, though that is an inference rather than a company statement.
Are AI features increasing SaaS seat growth?
Not in the disclosed numbers of these five vendors through mid 2026. Microsoft's seat growth stayed at 6%, Dropbox stayed flat and Zoom's enterprise net dollar expansion stayed just under 100%. AI shows up instead as separately counted add-ons, such as Copilot seats, and as faster revenue per user. Atlassian reports that AI-using customers expand seats slightly faster, which is the only direct link claimed.
Where to start
Pull your own seat count for your three largest productivity tools for the last six quarters and set it next to what you paid. If spend rose faster than seats, you are funding the pattern this post found in the filings.
Then ask each vendor one question at renewal: what share of this year's price increase is tied to features we have actually deployed? A vendor that can answer with usage data has a case. One that answers with a roadmap is pricing on the market, not on you.
Related analysis
The seat question sits inside a wider pricing shift. See how Workday hedged its seat model with flex credits, and why the market reacted the way it did in the SaaSpocalypse sell-off explained.
References
- Microsoft Investor Relations, FY26 Q4 metrics, July 2026, with earlier quarters' metrics pages. Used for seat growth, consumer subscribers and revenue growth.
- Microsoft, FY26 Q4 earnings release, 8-K exhibit 99.1, 29 July 2026. Used for the 30 million Copilot seats and adjusted revenue growth.
- Computerworld, Microsoft touts M365 Copilot momentum, claims 15M paid users, January 2026. Secondary coverage of the earnings call; used for 15 million and the 450 million base.
- Atlassian, Q4 FY26 shareholder letter, August 2026, plus the Q2 and Q3 FY26 letters and the FY26 10-K. Used for customer counts, cloud growth and seat expansion language.
- Zoom Communications, fiscal Q2 2027 results, 8-K exhibit 99.1, 25 August 2026, plus releases for the five prior quarters. Used for customer counts, churn and net dollar expansion.
- Dropbox, Q2 2026 results, 8-K exhibit 99.1, 6 August 2026, plus the Q3 2025, Q4 2025 and Q1 2026 releases and 10-Qs. Used for paying users, ARPU and ARR.
- Alphabet, Q1 2026 earnings release, 8-K exhibit 99.1, 29 April 2026, plus Q1 2025, Q3 2025 and Q4 2025 releases and the Q2 2026 call. Used for paid subscription totals and segment revenue.
The weakest thing about this source base: no vendor splits revenue growth into seats and price, so every "price did the work" reading here is an inference from two disclosed growth rates. Figures are current to filings released through August 2026.
Related reading