Microsoft Azure AI Backlog Hits $678 Billion, Up 84% Year-Over-Year

Microsoft's AI business posted eye-watering scale numbers. Azure's contracted backlog reached $678 billion, up 84% year-over-year, with guidance of 45% constant-currency growth through end of 2026. The commercial traction is broad: Microsoft 365 Copilot surpassed 30 million paid seats, GitHub Copilot revenue accelerated more than 60% quarter-over-quarter, and Azure AI Foundry now serves 100,000 customers. Underpinning it is roughly $175 billion in calendar-2026 capital expenditure — a staggering infrastructure bet.

The backlog figure is the standout: a $678 billion book of contracted-but-unrecognized revenue signals durable, multi-year demand and helps justify the capex. It also reflects Microsoft's dual strategy of selling both its own AI stack and OpenAI's models — GPT-6 Astra, Sol, and Luna all went live in Microsoft Foundry this week, with Microsoft executives emphasizing 'matching the right intelligence to the work' across the tiered lineup.

Wall Street responded: Stifel upgraded Microsoft from Hold to Buy with a $575 price target, citing Azure's 43% year-over-year revenue growth and Copilot momentum. The skeptical counterpoint is the capex-to-return question that hangs over the entire industry — $175 billion in annual spend requires sustained AI demand to pay off, and backlog is contracted intent, not booked cash. Still, of the hyperscalers, Microsoft's combination of enterprise distribution (Office, GitHub, Azure), the OpenAI partnership, and now a $678 billion backlog gives it arguably the most concrete AI-revenue story. Watch whether Copilot seat growth translates into per-seat expansion and whether Foundry's 100,000 customers deepen usage or remain experimental.