LOGO

OpenAI Microsoft Payment: Leaked Documents Reveal Costs

November 15, 2025
OpenAI Microsoft Payment: Leaked Documents Reveal Costs

Intensifying Financial Scrutiny at OpenAI

Following a period of significant investment and speculation surrounding a potential IPO, OpenAI is now facing increased financial oversight. Recently leaked documentation, reviewed by tech industry analyst Ed Zitron, offers insights into the company’s revenue streams and associated computational expenses over the last two years.

Revenue Sharing with Microsoft

Reports indicate that Microsoft remitted $493.8 million to OpenAI in revenue sharing payments during 2024. This figure experienced a substantial increase, reaching $865.8 million within the first nine months of 2025, as per the examined documents.

This arrangement stems from a prior agreement where Microsoft invested over $13 billion in the AI innovator, entitling them to 20% of OpenAI’s revenue. It’s important to note that neither organization has formally confirmed this percentage publicly.

Reciprocal Revenue Streams

However, the financial relationship is more nuanced. Microsoft also distributes revenue to OpenAI, approximately 20% of earnings generated from Bing and the Azure OpenAI Service. A source with knowledge of the situation confirmed this detail to TechCrunch.

Bing’s functionality is powered by OpenAI technology, and the Azure OpenAI Service provides developers and businesses with cloud-based access to OpenAI’s AI models.

The source clarified that the disclosed payments represent Microsoft’s net revenue share, excluding royalties received from Bing and Azure OpenAI. These figures are deducted from Microsoft’s internal revenue share calculations.

Estimating Microsoft’s Contributions

As Microsoft does not publicly detail the financial performance of Bing and Azure OpenAI, accurately determining the extent of their contributions remains challenging.

Nevertheless, the leaked information provides a valuable perspective on the financial dynamics of a leading private technology company – encompassing both revenue generation and expenditure.

Projected Revenue Figures

Based on the reported 20% revenue-sharing model, OpenAI’s revenue is estimated to have been at least $2.5 billion in 2024 and $4.33 billion during the first three quarters of 2025, though the actual figures are likely higher.

These estimates align with previous reports from The Information, which indicated OpenAI’s 2024 revenue around $4 billion and revenue for the first half of 2025 at $4.3 billion.

Sam Altman, OpenAI’s CEO, has stated that the company’s revenue surpasses reports of $13 billion annually, with an anticipated annualized revenue run rate exceeding $20 billion by year-end. He further suggested potential revenue reaching $100 billion by 2027.

Compute Costs and Infrastructure

Analysis of Zitron’s findings suggests OpenAI’s inference costs were approximately $3.8 billion in 2024. This expenditure rose to roughly $8.65 billion in the first nine months of 2025.

Inference refers to the computational resources required to operate a trained AI model and generate outputs.

Historically, OpenAI has primarily relied on Microsoft Azure for compute access, but has recently diversified its partnerships to include CoreWeave, Oracle, AWS, and Google Cloud.

Training vs. Inference Expenses

Previous reports estimated OpenAI’s total compute spend at $5.6 billion for 2024, with a “cost of revenue” of $2.5 billion for the first half of 2025.

A source familiar with the matter explained that OpenAI’s training expenses are largely non-cash, primarily covered by credits provided by Microsoft as part of their investment. Conversely, inference costs are predominantly cash-based. Training encompasses the resources needed to initially develop and refine an AI model.

Potential for Losses and Market Implications

These figures suggest that OpenAI’s inference costs may exceed its revenue, raising concerns about profitability.

This situation fuels ongoing discussions about the “AI bubble” and its potential implications for the broader AI industry. If a leading company like OpenAI is operating at a loss, it could signal challenges for other heavily-invested AI ventures.

Company Response and Contact Information

OpenAI declined to provide a comment on these findings. Microsoft did not respond to requests for comment from TechCrunch.

Do you have confidential information to share? TechCrunch is actively reporting on the AI industry, including company strategies and their impact. Contact Rebecca Bellan at rebecca.bellan@techcrunch.com or Russell Brandom at russell.brandom@techcrunch.com. Secure communication channels are available via Signal: @rebeccabellan.491 and russellbrandom.49.