Home » Leaked Documents Expose OpenAI’s Escalating Payments to Microsoft in AI Revenue Deal

Leaked Documents Expose OpenAI’s Escalating Payments to Microsoft in AI Revenue Deal

ByHardy Wayne
Leaked Documents Expose OpenAI's Escalating Payments to Microsoft in AI Revenue Deal

Unpacking the Financial Ties Between OpenAI and Microsoft

What does it cost to power the world’s most advanced AI models, and how much are tech giants like OpenAI willing to pay for the infrastructure behind them? Recent leaked documents offer a rare peek into the financial arrangements between OpenAI and Microsoft, highlighting the scale of revenue sharing and compute expenses in the rapidly evolving AI landscape.

Revenue Sharing Arrangements and Growth Projections

The documents indicate that Microsoft received $493.8 million in revenue share payments from OpenAI in 2024, a figure that surged to $865.8 million in the first three quarters of 2025. These payments stem from an agreement where OpenAI reportedly shares 20% of its revenue with Microsoft, tied to the software giant’s investments exceeding $13 billion in the AI firm. However, this percentage remains unconfirmed by either party, introducing some uncertainty to the exact terms.

  • Microsoft also shares approximately 20% of revenues from its Bing search engine—powered by OpenAI technology—and the Azure OpenAI Service, which provides cloud access to OpenAI’s models for developers and enterprises.
  • The leaked figures represent net revenue share, meaning they exclude Microsoft’s kickbacks to OpenAI from Bing and Azure, which are deducted internally. Without Microsoft’s detailed financial breakdowns for these segments, the full reciprocal flow remains opaque.
  • Based on the 20% share assumption, OpenAI’s revenue can be inferred at a minimum of $2.5 billion for 2024 and $4.33 billion for the first nine months of 2025, though actual figures are likely higher given prior estimates of $4 billion for full-year 2024 and $4.3 billion for the first half of 2025.

    • OpenAI CEO Sam Altman has publicly stated that the company’s annual revenue exceeds $13 billion and is on track to surpass $20 billion in annualized run rate by the end of 2025, with potential to reach $100 billion by 2027. These projections underscore the explosive market demand for AI services, but they also highlight the challenges of scaling amid intense competition from players like Google and Anthropic.

Surging Compute Costs and Their Broader Implications

A significant portion of the documents focuses on OpenAI’s compute expenditures, particularly for inference—the process of running trained AI models to generate responses—which has ballooned alongside revenue growth. Estimates suggest OpenAI spent approximately $3.8 billion on inference in 2024, rising to $8.65 billion in the first nine months of 2025.

  • Historically, OpenAI has relied heavily on Microsoft Azure for compute resources, though diversification efforts include partnerships with CoreWeave, Oracle, Amazon Web Services, and Google Cloud.
  • Training costs, which involve initial model development, are largely covered by non-cash credits from Microsoft’s investments, whereas inference expenses are predominantly cash outflows.
  • Prior analyses pegged OpenAI’s total compute spend at around $5.6 billion for 2024 and cost of revenue at $2.5 billion for the first half of 2025, indicating that inference alone may now outpace revenue in some periods.

    • These figures raise questions about profitability in the AI sector. If inference costs continue to exceed incoming revenue, as the data implies, it could signal sustainability challenges for high-valuation AI startups. The AI market, valued at over $200 billion globally in 2025 with projections to triple by 2030, is fueled by enterprise adoption—corporate AI spending is shifting paradigms, with budgets reallocating from traditional software to generative tools. Yet, this growth amplifies concerns over an “AI bubble,” where massive investments at valuations exceeding $150 billion for OpenAI alone may not yield immediate returns.

"OpenAI’s revenue is well more than reports of $13 billion a year," Altman noted recently, emphasizing the company's data center commitments nearing $1.4 trillion.

The interplay of revenue windfalls and escalating costs illustrates the high-stakes economics of AI development, where infrastructure demands could reshape market dynamics and investor confidence. As AI integrates deeper into business operations, how might these cost structures influence your organization’s adoption of generative tools—would optimizing for efficient inference become a priority in your strategic planning?

Similar Posts