For the better part of two years, the question hanging over Microsoft’s AI spending has been the one that follows every enterprise software bet: are people paying for it, and are they actually using it?
Wednesday’s fourth-quarter results gave Microsoft its clearest answer so far. Microsoft 365 Copilot has passed 30 million paid seats, with net seat additions more than doubling quarter-over-quarter, the sharpest jump the product has seen since launch, inside a quarter that saw Microsoft post $90 billion in revenue, up 18% year-on-year, and diluted earnings per share of $4.81, up 32% on a GAAP basis.
Azure Tops $100 Billion as Cloud Revenue Climbs 27%
The Copilot number sits on top of a broader cloud story. Microsoft Cloud revenue reached $59.3 billion for the quarter, up 27%, while commercial remaining performance obligation, a measure of contracted future revenue, grew 84% to $678 billion. Azure and other cloud services revenue rose 43% for the quarter, and for the full fiscal year Azure passed $100 billion in revenue for the first time.
Full-year revenue reached $331 billion, up 18%, with operating income up 21% to more than $155 billion. Amy Hood, chief financial officer, said premium-tier upgrades, including Copilot itself and the new E7 bundle, were driving the acceleration in Microsoft 365 Commercial cloud revenue this quarter.
Microsoft 365 Copilot Surpasses 30 Million Paid Seats
Satya Nadella, chairman and chief executive officer, gave the number straight to analysts on the call:
“We now have over 30 million paid Microsoft 365 Copilot seats.”
Large deployments are growing even faster than that headline figure. The number of customers buying more than 50,000 seats has grown more than sevenfold year-over-year, and the number of enterprise customers deploying Copilot to the majority of their workforce grew nearly 75% quarter-over-quarter. AstraZeneca, Boeing, Infosys, Koch, Procter & Gamble, Stellantis, Tata Consultancy Services, University of Pittsburgh Medical Center, Wells Fargo, and Wipro have each bought 60,000 or more seats.
Enterprise Deployments Deepen, From the NHS to Wall Street
Individual rollouts put flesh on those numbers. NHS England is extending Copilot to 505,000 clinicians and staff, the largest healthcare deployment of its kind, after a trial found it saved employees an average of 43 minutes a day. KPMG is expanding its rollout across a global workforce of more than 276,000 professionals, and HSBC has committed to 200,000 seats.
Microsoft’s new E7 suite, which bundles Copilot with E5, Entra, and Agent 365, has already been bought by hundreds of enterprise customers totalling millions of seats in its first two months on the market. EY’s deployment of E7 to 400,000 employees is Microsoft’s largest E7 win so far, distinct from the autonomous sourcing agent EY has built on Copilot, which is on track to handle more than 1,500 transactions over the next year.
Copilot’s Shift From Chat to Autonomous Agents
Copilot itself is changing shape. Microsoft made Copilot Cowork, its mode for chained, multi-step tasks, generally available last month, and this quarter introduced Autopilots: autonomous, long-running agents that include an always-on personal assistant Microsoft is calling OpenClaw. By the end of the quarter, Microsoft plans to fold chat, Cowork, Autopilots, and its coding assistant into a single product spanning consumer and commercial use.
Usage is catching up with that ambition. Conversations per user have nearly doubled year-over-year, and weekly engagement is now on par with Outlook and Teams, the two applications most organisations already run as switched-on-all-day infrastructure. Copilot matching that weekly cadence suggests it has moved past the experimental, novelty-driven usage that limited earlier rollouts and into the same category of habitual tool. Microsoft did not break out separate Teams user or revenue figures on the call, as is now standard practice, which makes this engagement comparison the clearest read available on how Copilot is embedding itself alongside Teams rather than sitting apart from it.
The time it takes a customer to reach high usage, meaning monthly active usage above 80% across its user base, has collapsed from months to days. Nadella adds:
“What used to be months is days from when a license is bought.”
User satisfaction scores have doubled over the past three quarters to an all-time high, while latency fell 25% this quarter alone.
Agent Governance Becomes a Bigger Story Than Seat Count
The seat and engagement numbers are only half the picture for IT and service management teams. Agent 365, the control plane Microsoft introduced to extend existing identity, security, and management frameworks to the agents organisations build on Copilot and Foundry, now has nearly 40 million agents registered across tens of thousands of companies, just two months after launch.
That scale is exactly the kind of agent sprawl we reported when Microsoft launched E7 and Agent 365 together, and it now sits alongside a similar shift lower down the stack. In Dynamics 365, usage-based credit consumption in customer service grew fourfold quarter-over-quarter, with customers including Northern Trust using it to drive more proactive service, a sign that the same seat-plus-consumption model reshaping Copilot is spreading into adjacent service management tools.
Analysts Say Results Show Progress, Not Proof
Not everyone reads the quarter as unambiguous vindication. Tracy Woo, principal analyst at Forrester, said the results point to Microsoft’s infrastructure spending starting to pay off, but stopped short of calling it validation of the company’s wider AI strategy.
“Microsoft’s strong revenue performance, combined with accelerating Copilot adoption, signals that its $190 billion data‑center buildout is beginning to deliver returns. But the results stop short of fully validating the company’s AI strategy. The new partnership with Anthropic helps reduce dependence on OpenAI, yet roughly 45 percent of commercial RPO remains tied to that single model provider. And while more AI products are scaling into enterprise‑grade workloads, the unresolved question is whether Microsoft’s infrastructure expansion can ultimately outrun the margin pressure that comes with supporting frontier‑model demand.”
Nadella addressed that dependence question himself when analysts pressed him on the call, saying Foundry now carries more than 11,000 models, including from OpenAI, Anthropic, Mistral, and xAI alongside Microsoft’s own MAI family, and that the number of customers building with models from more than one provider has grown fivefold since the start of the year. The Anthropic side of that story showed up in the numbers too: Microsoft booked a $3.2 billion gain during the quarter from its investment in the company, one of several items that pushed earnings per share above its own guidance.
