Microsoft reported a record quarter fueled by the Microsoft Azure cloud and the shift from talking about artificial intelligence (AI) to actually applying the technology at scale.

Azure revenue grew 24% year-over-year to reach $33 billion in Q2 2024, with 6 points of Azure cloud growth attributed to AI services.

Out of Microsoft’s 53,000 total Azure AI customers, more than a third are brand new Microsoft cloud customers. “By infusing AI across every layer of our tech stack, we are winning new customers and helping drive new benefits and productivity gains,” Microsoft CEO Satya Nadella said.

CFO Amy Hood similarly highlighted strong demand for Microsoft cloud and AI services, which drove better-than-anticipated growth in large, long-term Azure contracts.

Operating in a cloud environment has been instrumental in building AI, but “now AI is redefining what the cloud looks like, both at the infrastructure level and the app model,” Nadella said.

To that point, generative AI (genAI) will have a foundational impact on the tech stack as most enterprises currently know it, particularly as enterprises begin deploying AI at scale.

With genAI, the core compute architecture itself needs to change – everything from power density and data center design to the main compute unit, the network and memory architectures. “As the core computer architecture changes, I think every workload changes,” Nadella said.

“The most exciting thing for me in the last year has been to see how our data layer has evolved to be built for AI,” he told investors.

Microsoft’s AI Fabric, for example, separates storage and compute layers. “In compute, we’ll have traditional sequels that have [Apache] Spark, and you can have an Azure AI job on top of the same data lake,” Nadella said. “Then the business model – you can combine all of those different compute [layers]. That’s the type of compute architecture” that Microsoft expects will gain traction.

As executives touted the surge in demand for Azure cloud and AI services, investors showed concerns about the possibility of the hyperscaler being constrained by GPU supplies and other scale issues.

Hood, however, feels “really good” about Microsoft’s efforts to add data center capacity. “You started to see the acceleration in our capital expense starting almost a year ago, and you've seen us scale through that process,” she told investors.

Microsoft’s investment in servers and expanded data center footprints – including third-party data center capacity– will help it meet the evolving demand going forward. “Obviously, [Azure] is being built by us, but we’ve also used third-party capacity to help when we could have that help us in terms of meeting customer demand,” she said.

The company does plan to continue to increase its data center capex looking forward, “given what we see in terms of pipeline.”