Microsoft is continuing its trend of standing up 1GW of capacity every three months, having managed the task for the second consecutive quarter.

CEO Satya Nadella revealed that the company's capacity expansion plans remain "on track to double our overall footprint in just two years" during the company's Q3 FY2026 earnings call held on April 29.

Capex for the quarter, meanwhile, was $31.9 billion, two-thirds of which went on short-lived assets like GPUs and CPUs, and the remainder towards data centers. This was less than the prior quarter's $37.5bn, though during that earnings call, CFO Amy Hood had said to expect a lower capex in Q3 of FY2026.

Furthermore, $4.7bn was spent on data center leases, similarly down from the previous quarter's $6.7bn.

This comes after reports that the company was ramping up its efforts to grab capacity, having previously stepped back on some projects. In March last year, the tech giant was reported to be canceling or delaying several projects - at least 2GW - globally.

According to the Information, this followed CFO Amy Hood's decision to put a curb on spending when, at the end of 2024, it started to exceed the $80bn budget. The publication reported that the projects left by Microsoft were rapidly taken up by competitors "to an extent not previously reported," and the company is now having to play catch-up. Microsoft has maintained that it is set to be capacity-constrained through the end of this fiscal year, at least.

However, Microsoft has said that it is expecting Q4 to have a capex of $40bn, and full year spend to reach $190bn, in part due to a "$25bn higher component pricing."

While competitors Amazon and Google stated that, for the most part, they were not expecting to be heavily impacted by the global supply chain issues that have led to component costs rising - particularly memory, CFO Hood noted that "even with these additional investments and continued efforts to bring GPU, CPU, and storage capacity online faster, we expect to remain constrained at least through 2026."

The company has made a slew of investment commitments over the last few months, including several in the last month alone, such as $10bn to expand infrastructure in Japan, $18bn in Australia, $5.5bn in Singapore, and a multi-billion-dollar investment in Ontario, Canada.

While constraints remain, Microsoft's cloud revenue for Q3 was $54.5bn, up 29 percent Year-on-Year (YoY), and up from the previous quarter's $51.5bn, which itself was the first time the cloud business had surpassed $50bn. Azure revenue, meanwhile, grew 40 percent in Q3, a slight increase from growth in Q2.

Remaining performance obligation increased to $627bn this quarter, up 99 percent YoY, and with an average duration of two and a half years. 25 percent of this is expected to be recognized in the next 12 months.

Hood also noted that the company is seeing higher margins for its AI business than was perhaps expected. "We have been talking about where this AI business of ours has been in the cycle compared to the cycle we saw with the cloud, which now seems very long ago, and how margins were actually better and have remained better in our AI business versus where we saw them in the cloud transition looking back."

She added that the company is also working hard on its first-party infrastructure stack and focusing on "efficiency work," explaining: "We have been in an accelerated phase of trying to get as much capacity as we can into production."

She later elaborated: "When we talk about some acceleration into what I would call the first half of FY27—the second half of the calendar year—it means we are getting insights into our abilities to increasingly put pressure on efficiencies, speed up the deliveries into our data centers, and make that revenue-ready as quickly as we can.

"I would expect the pressure between first-party usage and being able to meet Azure demand will persist, as I said, but we are doing our best to get things in as quickly as we can—hence the capex number that we see in the second half of the year."

Unavoidable in the earnings call was the mention of the recently changed terms of Microsoft's agreement with OpenAI. Announced earlier this week, the new terms see Microsoft losing its IP exclusivity rights, but the company will continue to receive revenue share payments "subject to a total cap" through to 2030, and these payments are no longer subject to OpenAI achieving Artificial General Intelligence (AGI). Microsoft will also no longer pay a revenue share to OpenAI.

CEO Satya Nadella told analysts: "Overall, we feel good about our partnership with OpenAI. I am always very focused on any partnership and ensuring that there is a win-win construct at all times; that is how you can remain good partners. In this case, it starts with IP. Amy referenced this. We have a frontier model royalty-free with all the IP rights that we will have access to all the way to '32, and we fully plan to exploit it .... They are a large customer of ours, not just on the AI accelerator side, but also on all the other compute side, and we want to serve them well. And then, of course, we have our equity. Overall, I think the construct, as they have grown and we have grown, and our customers also have different expectations in terms of their model diversity, has evolved, but I feel very good about where we are."

Following the earnings call, Microsoft's share price dropped more than six percent.