Microsoft — Holding
Claim reviewed: Microsoft reported 15 million paid Microsoft 365 Copilot seats in FY2026 Q2 (28 Jan 2026). The number of customers with more than 35,000 Copilot seats tripled year-over-year. Agent 365 was introduced as an enterprise governance/identity/security control plane for agents.
Reviewed 55 days after initial log (2026-04-19). Original claim has three components, each re-verified.
15 million paid Microsoft 365 Copilot seats at FY26 Q2 close. Confirmed in the Q2 press release and the earnings call transcript of 28 Jan 2026. The figure is paired with a +160% year-over-year seat-add growth rate. No correction, restatement, or clarification has appeared in the 86 days since disclosure.
Customers with >35,000 Copilot seats tripled year-over-year. Confirmed in the same earnings disclosure, with named customers in the Q2 cohort: Publicis (95,000+ seats), plus Coca-Cola, Fiserv, ING, NASA, PepsiCo, University of Kentucky, University of Manchester, US Department of the Interior, and Westpac. The named-customer list is the strongest corroboration. It is unusual for a vendor to enumerate this many 35k+ seat accounts unless each is individually documented internally.
Agent 365 introduced as an enterprise governance, identity, security, and management control plane for agents. Confirmed in the Q2 announcement cycle. Positioning language from the original disclosure, “extend their existing governance, identity, security, and management to agents” across Microsoft cloud and any other cloud, is carried unchanged in downstream coverage through April 2026.
Counter-considerations weighed but not persuasive:
- Paid vs. trialling seats. Microsoft uses the word “paid” explicitly; analyst estimates (Bank of America: roughly 3.5% of the M365 commercial installed base) triangulate to a similar absolute seat count. The disclosure wording is tight.
- “Seats” vs. “active users.” The 15M figure is seats (entitlements sold), not monthly active users. The claim as logged specifies “paid seats”, so the wording is accurate; interpretation as usage would be a misread, not a claim failure.
- Shelfware risk. Commentary from Neowin and Alphastreet through Q1 2026 has raised the question of whether seat counts translate to sustained usage. That is a secondary question about ROI, not about whether the 15M figure is accurate. It does not undermine the claim as stated.
Next checkpoint: 2026-05-24 (30 days). Microsoft FY26 Q3 reports on 29 Apr 2026 (five days after this review). That release will update seat count, large-customer cohort, and any Agent 365 traction, and should drive a clearer strengthened or weakened verdict at the next review window.