Microsoft investors sweat cloud giant’s OpenAI exposure

Steven

Microsoft investors sweat cloud giant’s OpenAI exposure

Summary

Microsoft posted a strong Q2 — profits rose 60% YoY to $38.5bn on $81.3bn revenue — but investors reacted nervously, knocking the share price down about 6% after hours. The cause: heavy concentration of future revenue and commitments tied to AI model developers. Roughly 45% of Microsoft’s $625bn backlog is attributable to OpenAI, which committed to an incremental $250bn of Azure services as part of a restructuring. Anthropic also accounts for a large chunk via a $30bn Azure commitment. Management says the remainder of the backlog is diversified and largely contracted, but concerns persist because the AI customers driving the backlog have yet to demonstrate consistent profitability and demand expensive, fast-depreciating GPU capacity.

Key Points

  • Microsoft’s Q2: profits +60% YoY ($38.5bn) on $81.3bn revenue, yet shares fell ~6% after hours.
  • About 45% of Microsoft’s $625bn backlog is linked to OpenAI after a $250bn Azure purchase commitment.
  • Anthropic’s $30bn Azure commitment also inflates backlog and requires substantial Nvidia-based GPU deployments.
  • Q2 capital expenditure topped $37.5bn, mostly on GPUs/CPUs that depreciate quickly.
  • Management argues most capacity is contracted for its useful life, reducing some risk — but customer payment and profitability risk remain.
  • Microsoft guides Q3 revenue of $80.65–$81.75bn (+15–17% YoY) and expects lower capex versus Q2.

Content Summary

The article explains that despite headline earnings strength, investor unease centres on Microsoft’s concentrated exposure to a few large AI model vendors and the huge infrastructure spend required to support them. OpenAI and Anthropic commitments make up a material share of Microsoft’s performance obligations, and delivering on those commitments requires large GPU investments (including Nvidia Grace Blackwell and Vera Rubin systems). CFO Amy Hood emphasised that a significant portion of capacity is contracted for its useful life, but the piece warns that signed deals don’t eliminate the risk of customers failing to pay or AI startups failing to turn a profit.

Context and Relevance

This matters because it highlights a major dynamic in the cloud and AI ecosystem: growth is being driven by a small number of hyperscale model developers, which concentrates commercial risk even as it fuels revenue growth. For CIOs, investors and cloud customers, the story signals potential volatility in pricing, capital allocation and strategic partnerships — and it underlines the growing importance of Nvidia and other accelerator suppliers in the AI supply chain. It also ties into broader trends: enterprise caution around AI agents, ongoing datacentre expansion and environmental/resource considerations, and the evolving economics of cloud compute.

Why should I read this?

Look — headline profits are great, but if you want to understand where Microsoft’s growth (and risk) really sits, this is the short read you need. It shows how a handful of AI customers can tilt a giant’s fortunes, why GPUs are now a balance-sheet issue, and what investors are worried about next. Worth five minutes if you follow cloud economics or AI infrastructure.

Source

Source: https://go.theregister.com/feed/www.theregister.com/2026/01/29/microsoft_earnings_q2_2026/