Newly reported leaked documents are giving a sharper, if still incomplete, look at the money moving between OpenAI and Microsoft. The figures point to large revenue-share payments, rising inference costs, and a business model under growing scrutiny as AI investment continues to accelerate.
What the leaked documents show
Tech blogger Ed Zitron reported that Microsoft received $493.8 million in revenue share payments from OpenAI in 2024. In the first three quarters of 2025, the figure rose to $865.8 million, according to documents he viewed.
The payments matter because OpenAI reportedly shares 20% of its revenue with Microsoft under an earlier deal tied to Microsoft’s investment of over $13 billion in the AI startup. The source article notes that neither OpenAI nor Microsoft has publicly confirmed that percentage.
On its own, the leaked payment number might appear to offer a simple way to estimate OpenAI revenue. But the arrangement is more complicated than a one-way flow from OpenAI to Microsoft.
Why the Microsoft relationship is not simple
According to a source familiar with the matter cited by TechCrunch, Microsoft also shares revenue with OpenAI. That includes about 20% of revenue from Bing and Azure OpenAI Service.
Bing is powered by OpenAI, while Azure OpenAI Service sells cloud access to OpenAI models to developers and businesses. Those products create revenue streams that can flow back from Microsoft to OpenAI.
The same source told TechCrunch that the leaked payments refer to Microsoft’s net revenue share, not gross revenue share. In plain terms, the figures do not include whatever Microsoft paid to OpenAI from Bing and Azure OpenAI royalties. Microsoft deducts those amounts from its internally reported revenue-share numbers, according to this person.
That distinction is important. Microsoft does not break out how much it makes from Bing and Azure OpenAI in its financial statements, so there is no clear public way to calculate the full amount Microsoft is returning to OpenAI through those channels.
What the figures imply about OpenAI revenue
Using the widely reported 20% revenue-share figure, the leaked payments imply that OpenAI revenue was at least $2.5 billion in 2024 and $4.33 billion in the first three quarters of 2025. TechCrunch noted that the real number is very likely higher because the leaked payments are described as net revenue share.
Other reported figures help frame the scale. Previous reports from The Information put OpenAI’s 2024 revenue at around $4 billion and its revenue from the first half of 2025 at $4.3 billion.
Sam Altman has also recently said OpenAI’s revenue is “well more” than reports of $13 billion a year. He said the company will end the year above $20 billion in annualized revenue run rate, which TechCrunch described as a projection rather than guidance on actual revenue. He also said the company could even hit $100 billion by 2027.
Taken together, the reported numbers show a company with fast-rising revenue and a close financial link to Microsoft. They also show why outside observers are paying attention not only to sales growth, but to the cost of delivering the product.
Inference costs are the harder question
The same leaked materials, through Zitron’s analysis, suggest OpenAI may have spent roughly $3.8 billion on inference in 2024. That figure increased to roughly $8.65 billion in the first nine months of 2025.
Inference is the compute used to run a trained AI model so it can generate responses. For a company serving many users and businesses, inference is not a side cost. It is part of delivering the core service every time the model is used.
OpenAI has historically almost exclusively relied on Microsoft Azure for compute access. The company has also struck deals with CoreWeave and Oracle, and more recently with AWS and Google Cloud.
Previous reports put OpenAI’s entire compute spend at roughly $5.6 billion for 2024 and its “cost of revenue” at $2.5 billion for the first half of 2025. TechCrunch also reported that a source familiar with the matter said OpenAI’s training spend is mostly non-cash, meaning it is paid by credits Microsoft awarded OpenAI as part of its investment. By contrast, the source said inference spend is largely cash.
Why this fuels the AI economics debate
The leaked documents do not provide a full financial picture of OpenAI. They do, however, sharpen the central question around generative AI: how much revenue can these services produce compared with the cash needed to run them?
TechCrunch noted that the numbers imply OpenAI could be spending more on inference costs than it is earning in revenue. That possibility is significant because OpenAI is one of the most closely watched companies in the private markets.
If OpenAI is still spending heavily to run its models, the implications reach beyond one company. The figures add to debate over whether current AI investments and valuations can be supported by the economics of model deployment.
OpenAI declined to comment to TechCrunch. Microsoft did not respond to TechCrunch’s request for comment.