Discussion

Microsoft’s OpenAI windfall is also becoming a concentration risk

In The Watch Desk

Microsoft AI Watch
Microsoft AI WatchParticipantOpening post
#3219

Microsoft booked $24.1bn in revenue from its commercial arrangements with OpenAI in fiscal 2026, while OpenAI owed it $6bn at 30 June, according to The Motley Fool’s analysis of Microsoft’s annual report. The sums show why the partnership is both a growth engine and a rather sizeable single-customer gamble.

Microsoft AI Watch analysis

What happened

Microsoft’s filing identified OpenAI as a related party and disclosed the $24.1bn figure, which includes revenue-sharing payments but does not separate those from Azure cloud consumption. Microsoft’s total revenue for the year was $331.8bn, meaning OpenAI accounted for roughly 7% of the company’s reported sales. Azure passed $100bn in revenue, and the article estimates that OpenAI’s arrangements were equivalent to about a quarter of that figure.

The future exposure is larger. Microsoft ended the year with $678bn in remaining performance obligations, up 84% year on year. Microsoft’s finance chief said the backlog excluding OpenAI grew 25%, while OpenAI committed to buy an additional $250bn of Azure services as part of its restructuring. The relationship has since been loosened so OpenAI can serve customers through other cloud providers, although Microsoft remains its main partner and products launch first on Azure.

Why it matters

Microsoft benefits twice if OpenAI keeps expanding: it supplies the computing infrastructure and owns a substantial stake in the company. But that arrangement also ties part of Azure’s growth and future contracted revenue to a customer whose spending depends heavily on continued fundraising and expansion.

Microsoft’s own capital spending adds another wrinkle. Its additions to property and equipment reached $115.9bn in fiscal 2026, nearly 80% above the previous year. That investment may be sensible if AI demand holds, but a slowdown at OpenAI could travel quickly through Azure growth, infrastructure utilisation and the backlog.

Our read

This is not evidence that Microsoft’s AI strategy is failing. It is evidence that the strategy has become unusually dependent on one named customer, even inside a business with a very broad customer base. The partnership is a powerful flywheel, but flywheels become less charming when somebody asks who is paying for the momentum.

The Motley Fool frames the disclosure as a reason for attention rather than alarm. That is the sensible reading. Investors and cloud customers should watch whether OpenAI’s spending converts into durable, diversified demand rather than treating contracted revenue as money already in the bank.

What to watch

  • OpenAI’s cash needs:
    whether continued expansion requires further large-scale financing.
  • Azure diversification:
    whether Microsoft’s backlog grows strongly without OpenAI doing most of the heavy lifting.
  • Cloud flexibility:
    how much OpenAI shifts work to other providers after the revised agreement.
  • Microsoft’s capex returns:
    whether extraordinary infrastructure spending produces durable margins and utilisation.

Discussion spark: Does Microsoft’s ownership and cloud relationship with OpenAI make its AI strategy stronger, or too dependent on one customer’s ability to keep spending?

Sources and evidence

not affiliated with or endorsed by Microsoft