- Azure finally crossed the $100B mark, with growth hitting its fastest clip in four years
- Microsoft is pouring billions into data centers, leases and compute capacity — and still can’t keep up with demand
- The big question now is whether Microsoft can turn its AI infrastructure land grab into real profitability before capex outruns returns
Microsoft’s cloud engine positively purred in its fiscal Q4 2026, with Azure growth rebounding and propelling the business beyond a key revenue milestone. But the continued upward creep of its infrastructure bill raised red flags with analysts.
Microsoft doesn’t usually break out Azure revenue. But the company noted that Azure and other cloud services revenue rose 43% year on year, with standalone Azure revenue topping $100 billion for the first time ever in its full fiscal year 2026 (July 1, 2025 to June 30, 2026).
Azure growth was the highest it’s been since Q3 2022, when it came in at 46%.
Consolidated revenue of $90 billion was up 18% year on year, with net income increasing 31% to $35.8 billion. Net income included a $3.2 billion boost from the company’s investment in Anthropic.
Intelligent Cloud revenue, which includes server products and cloud services (including Azure) as well as enterprise and partner services, rose 32% to $39.3 billion, driven by the increase in Azure revenue.
Continued capacity imbalance
The jump in Azure revenue was at least partially attributed to the staggering pace at which Microsoft is adding compute capacity.
“We added 31 new datacenters across 5 continents this quarter, bringing the total to 88 this year, as we expand our footprint in response to accelerating demand,” CEO Satya Nadella said on the company’s earnings call. “All up, we added another gigawatt of capacity this quarter and remain on track to roughly double our overall capacity in just two years.”
CFO Amy Hood noted that capacity added in the quarter was “quickly monetized.”
But it still can’t keep up with demand. The company’s order backlog climbed to $678 billion, up 84% year on year (or 25% if you exclude its massive OpenAI deals). Around 30% of this backlog – or $203.4 billion – is expected to convert to revenue in the next 12 months.
Microsoft looks to be trying to quicken the pace of adding new capacity, adding $132.5 billion worth of leases – primarily for data centers – in the quarter. All told, it’s pipeline of leases not yet commenced stands at $329.1 billion, with those lease terms set to begin between its fiscal 2027 and fiscal 2033.
Capex was $41 billion in the quarter, and Microsoft paid $5.6 billion in finance leases, mostly for data center sites. Hood said Microsoft is looking to shift more of its data center leases from finance leases to operating leases, which means some of this expense will drop off from its capex line.
In its fiscal Q1 (calendar Q3), Microsoft expects to spend more than $50 billion in capex.
Microsoft's results warrant close scrutiny
Sid Nag, Tekonyx Founder and Chief Research Officer, noted on LinkedIn that Microsoft’s results warrant close scrutiny.
In particular, he pointed to the relationship between capex – which is expected to come in at $175 billion for calendar 2026 and the Azure revenue Microsoft disclosed. Nag noted “expected 2026 capex is roughly 1.75x Azure’s current annual revenue,” highlighting “the unprecedented capital intensity behind Microsoft’s AI strategy.”
“Azure demand is clearly accelerating, but Microsoft is deploying capital even faster than it is proving attractive AI returns,” Tekonyx Founder and Chief Research Officer Sid Nag wrote. “The company has demonstrated scale. It has not yet demonstrated AI infrastructure profitability, capital efficiency or return on invested capital.”
Addressing analyst questions about whether the industry is potentially building an oversupply of data centers and capacity, Hood noted that short-lived assets like CPUs and GPUs are a key cost driver.
“So, if the demand environment changes, you just slow down what is, in fact, the largest component and the driver of COGS [cost of goods sold],” she said. “The investment into land and data center builds is actually quite flexible. It’s a smaller percentage of the overall cost structure, and timing can be changed on much of that.”
Read the latest about Microsoft
Microsoft taps Mistral for more AI models, European compute capacity
Microsoft launches Frontier Company with $2.5B for scaling enterprise AI transformation