New Disclosures Show OpenAI Drives the Majority of Microsoft AI Revenue Growth

New financial disclosures reveal that the vast majority of Microsoft’s generative AI sales are driven directly by its partnership with OpenAI, rather than its own proprietary Copilot products. The revelation, first reported by Bloomberg, exposes a stark commercial reality: Redmond’s massive infrastructure bet is currently paying off through its proxy, not its own software. For investors expecting Microsoft to dominate the enterprise SaaS layer with its own models, this is a sobering course correction.

The Reality Behind Microsoft AI Revenue

For the past two years, Microsoft CEO Satya Nadella has pitched a unified narrative: Azure is the world's AI supercomputer, and Copilot is the ubiquitous AI companion transforming knowledge work. However, the underlying numbers paint a more lopsided picture. According to the disclosures, the consumption of OpenAI’s API services on Azure accounts for the lion's share of Microsoft’s cloud-based AI growth, leaving Microsoft’s in-house offerings as secondary contributors to the bottom line.

This lopsided distribution of Microsoft AI revenue raises critical questions about the software giant's long-term ROI. Microsoft has committed over $13 billion to OpenAI and is spending tens of billions more on data center capital expenditures. If the primary driver of this infrastructure spend is third-party demand for OpenAI's models rather than Microsoft's high-margin Copilot subscriptions, the economic calculus of Redmond's AI strategy shifts dramatically.

Is Copilot Underperforming Its Hype?

The enterprise adoption of Microsoft Copilot has been met with mixed signals. While the company frequently touts seat-license growth in earnings calls, anecdotal reports of high churn and low daily active usage have persisted. Enterprise IT buyers are increasingly questioning the $30-per-user monthly premium for what is often perceived as an inconsistent productivity wrapper. Meanwhile, developers and enterprises building proprietary tools are choosing to bypass Copilot entirely, opting instead to build directly on OpenAI's GPT-4o via Azure's API.

This dynamic creates a strategic vulnerability for Microsoft. While it captures a portion of OpenAI's upside through cloud hosting fees and its equity stake, it remains highly dependent on a volatile partner led by OpenAI CEO Sam Altman. If OpenAI decides to further diversify its cloud hosting to other providers like Oracle, or build its own infrastructure, Microsoft’s primary engine of AI growth could be severely compromised.

The Shifting Balance of Power

This disclosure shifts the commercial balance of power back toward OpenAI. It proves that the value in the generative AI stack is still concentrated at the frontier model layer, rather than the application layer that Microsoft hoped to monopolize. For founders and investors, the lesson is clear: building wrappers around someone else's intelligence is a tough sell when enterprises would rather access the raw engine directly.

Ultimately, Microsoft is acting as an expensive landlord for OpenAI's tenants. Unless Microsoft can successfully convince the enterprise that its proprietary Copilot software adds indispensable value beyond the raw API, its positioning in the AI race looks less like a sovereign empire and more like a highly leveraged real estate play.