Although Microsoft exceeded analyst expectations in its latest quarterly results for FY26 Q2, posting more than $50 billion in cloud revenue for the first time — a 26% year-over-year increase — the company’s share price fell by double digits during the following trading session. Management emphasized that its AI business is now larger than some of its decades-old core products. Still, investors appear unconvinced.

The sell-off reflects growing concern that Microsoft’s soaring AI infrastructure spending may not be translating into proportional revenue growth. The company invested $37.5 billion in the quarter, with roughly two-thirds allocated to GPUs and CPUs. At the same time, cloud gross margin declined to 67%, and guidance for the next quarter points to 65%, while Azure growth is expected to slow from 39% to 37–38%.

For investors, this combination is a warning signal: record capital expenditures, slowing growth, and shrinking margins. Microsoft says most GPU purchases are already contractually committed for their full useful life, underscoring how locked-in these investments have become.

CFO Amy Hood argued that Azure could grow faster if more compute capacity were available. Instead, Microsoft is strategically allocating scarce GPUs: first to internal products such as Microsoft 365 Copilot and GitHub Copilot, then to R&D teams, and only afterward to Azure customers.

Copilot revenue grew more than 160% year over year, with the number of customers holding over 35,000 licenses tripling and average conversations per user doubling. Microsoft 365 Copilot now has 15 million paying users. However, these figures still do not clarify whether Copilot’s performance justifies Microsoft’s enormous AI investments — especially compared to rivals like ChatGPT.

Microsoft’s cloud growth hinges heavily on OpenAI

Investor concerns are further amplified by Microsoft’s deep dependency on OpenAI. According to the earnings report, 45% of Microsoft’s $625 billion commercial backlog comes from a single customer: OpenAI. Large contracts with OpenAI and Anthropic alone pushed booking growth to 230%, reinforcing fears of a circular AI economy in which cloud providers and model startups inflate each other’s revenues.

Pressed on this issue, Hood deflected, highlighting that 55% — about $350 billion — comes from a diversified customer base, spanning industries, regions, and Azure services, which grew 28% year over year.

Regarding the OpenAI partnership itself, Hood offered only broad praise, calling it a “great partnership” and reaffirming Microsoft’s role as the scaling provider — without directly addressing the risks of such heavy reliance on a single client.

Conclusion: Microsoft’s AI strategy is reshaping its business at unprecedented speed, but rising costs, slowing cloud growth, and deep dependence on OpenAI are testing investor confidence. The company now faces mounting pressure to prove that its massive AI investments can deliver sustainable, diversified returns.