Microsoft reported fiscal Q2 2026 earnings after the bell on Thursday, and the results confirmed what analysts had been anticipating for months: the company’s Azure cloud platform has become, in measurable terms, an AI infrastructure business with a productivity software side. Azure revenue grew 34% year-over-year in constant currency — the fastest quarterly growth rate since Q3 2022 — with AI services accounting for 18 percentage points of that growth, up from 12 points in Q2 2025.
Copilot: The Enterprise Adoption Number That Matters
The headline figure from CFO Amy Hood’s call with analysts was the Copilot seat count: Microsoft Copilot for Microsoft 365 has now been activated by over 100 million enterprise users across paying tenants. That is not the same as 100 million daily active users — enterprise software rarely achieves that — but the activation number represents organizational commitment, procurement decisions, and IT provisioning. It is the funnel metric, and it has expanded fivefold in 12 months.
Microsoft declined to break out precise daily active user numbers, but noted that customers using Copilot features at least weekly generate 2.4x the revenue per seat compared to baseline Microsoft 365 E3 subscribers. The company also disclosed that Copilot-assisted code completions in GitHub Copilot now account for 40% of all code committed through the platform — a figure that aligns closely with Google’s recently disclosed internal benchmark of 75% AI-assisted code contribution, though the methodologies differ.
“We are past the ‘is this real?’ question in enterprise AI,” said CEO Satya Nadella on the earnings call. “The questions now are about scale, customization, and integration depth — and those are problems we are structured to solve.”
Azure AI Infrastructure: Where the Margin Is
Azure AI services — encompassing model hosting through Azure OpenAI Service, Foundry platform charges, and AI search — generated an estimated $8.2 billion in annualized revenue based on Q2 run rate, according to analyst estimates from MoffettNathanson and Bernstein Research. Microsoft does not break this out as a separate segment, which has become a mild frustration for institutional investors trying to value the AI business independently.
The Azure OpenAI Service alone hosts requests for over 30,000 enterprise organizations, Microsoft disclosed — a figure that includes Fortune 500 companies, government agencies, and mid-market firms. Average contract value for Azure OpenAI Service enterprise agreements has risen roughly 40% year-over-year as customers shift from experimental to production deployments running at scale.
Capital expenditure in Q2 reached $22.6 billion, almost exclusively directed at AI datacenter infrastructure — servers, networking, and leased facility expansion. Microsoft is the single largest customer of Nvidia’s Blackwell-family GPUs by contracted volume, according to supply chain sources cited by Bernstein Research. The company is also expanding its AMD Instinct MI350 deployment for inference workloads where cost per token takes priority over raw throughput.
The Competitive Pressure Building Below
The strong Azure numbers land in a market where Google Cloud and AWS are also reporting AI-driven acceleration. Google Cloud grew 29% year-over-year in its own Q2, with Gemini API consumption described by Google CFO Anat Ashkenazi as ‘ahead of our most optimistic internal projections.’ AWS re:Invent 2025 commitments around Trainium and Inferentia chips are beginning to show in customer workloads, though AWS’s AI-specific revenue disclosure remains limited.
The more immediate threat Microsoft is navigating is enterprise AI commoditization: as more organizations build their own fine-tuned models on open-weight foundations like Meta’s Llama 4 or Mistral Large 3, the marginal value of Azure OpenAI Service — which bundles access to GPT-4o and o3 model families — may compress. Microsoft’s response is to position Azure AI Foundry as the orchestration and agent deployment layer, rather than defending any single model’s premium.
The Q2 results suggest that strategy is working, at least for now. Operating income grew 25% year-over-year to $31.4 billion, with the Intelligent Cloud segment — home to Azure — contributing $17.1 billion of that at a 44% operating margin. AI is, at this point in the cycle, not just a growth story for Microsoft. It is the margin story.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.