Microsoft Azure Grows 43% as AI Demand Exceeds Supply; Nadella Says 'Every Model Is Substitutable'

Microsoft's fiscal results underscored both the scale of AI demand and a strategic thesis about model commoditization. Azure grew 43% year-over-year, Microsoft Cloud reached $214.4 billion for the fiscal year, and remaining performance obligations — a proxy for contracted future revenue — climbed to $678 billion. Management reiterated that demand for AI capacity continues to exceed supply, a constraint echoed across the hyperscalers this week.

The most quoted line was Nadella's: 'every model is substitutable.' It captures Microsoft's push toward multi-model architectures, where customers mix and match models by task and cost rather than committing to a single provider — a framing that conveniently supports Microsoft's parallel move to route commodity Copilot traffic to its own MAI models while keeping OpenAI for premium workloads.

Competitive context: the 'substitutable' thesis is a direct challenge to the idea that any one lab holds a durable moat, and it dovetails with the week's price war (AWS cutting GPT-5.6 prices, DeepSeek and Qwen undercutting on cost). If models are interchangeable commodities, the value accrues to the cloud layer, distribution, and enterprise integration — where Microsoft is strong.

Caveats: 'substitutable' understates real quality gaps on hard tasks, and Microsoft's own reliance on OpenAI for frontier work belies the commodity framing at the top end. Microsoft also faces a UK CMA probe over Microsoft 365 subscription changes, a regulatory overhang. What to watch: whether capacity constraints ease, how much Copilot traffic actually shifts to MAI, and whether enterprises genuinely treat models as swappable in production.