Podcast Summary: The Last Invention is AI

Episode: Industry Shocked After Leak Shows OpenAI Losing Money to Microsoft Fees
Date: November 18, 2025
Host: The Last Invention is AI

Episode Overview

This episode dives into recently leaked documents revealing the financial details between OpenAI and Microsoft, focusing on the staggering compute costs and revenue sharing arrangement. The discussion analyzes what these leaks suggest about OpenAI's real financial health, their relationship with Microsoft, and the broader implications for the artificial intelligence industry.

Key Discussion Points & Insights

1. The Leaked Financials: OpenAI’s Massive Payments to Microsoft

• Microsoft's Revenue Share Receipts
  • In 2024, Microsoft received $493.8 million from OpenAI as part of their revenue share agreement, which nearly doubled to $865.8 million (first three quarters of 2025).
    • “So according to his report, he said that in 2024 Microsoft received $493.8 million in revenue share payments from OpenAI… that number jumped to 865.8 million. And that's just in the first three quarters of this year.” – Host [01:10]
  • Projected payments for 2025 are expected to exceed $1 billion.
• Origin of the Revenue Share
  • Following Microsoft’s $10 billion investment (and $13 billion in total funding), they agreed to a 20% revenue share from OpenAI.
  • Although neither company has officially confirmed the 20% figure, this has been widely reported.
    • “So 20% of the revenue OpenAI made, they were going to pay to Microsoft. Neither people have also publicly confirmed that percentage of the 20% by the way, but this has been reported and leaked by a lot of different people.” – Host [02:31]

2. OpenAI–Microsoft: The Partnership Structure

• Mutual Revenue Sharing
  • Not only does OpenAI pay 20% of its revenue to Microsoft, but Microsoft reportedly shares 20% of revenue back to OpenAI for Azure OpenAI Services and Bing integration.
• Opaque Financial Reporting
  • Microsoft does not break out Bing or Azure OpenAI revenues in public statements, making true profitability hard to gauge.
    • “So it's pretty hard to get a good estimate on how much they are currently kicking back and all of this.” – Host [04:02]

3. Estimating OpenAI's True Revenue and Costs

• Using the 20% Share to Approximate OpenAI’s Revenue
  • Leaks suggest OpenAI’s 2024 revenue was at least $2.5 billion, rising to $4.33 billion in the first three quarters of 2025.
• Conflicting Reports
  • Some sources estimate a 2024 revenue of “about $4 billion” and $4.3 billion in the first half of 2025, though these numbers slightly conflict but are in the same ballpark.
    • “There are some other reports that came out of the information that said that OpenAI's 2024 revenue was about 4 billion and that its revenue from the first half of this year was 4.3 billion.” – Host [05:05]

4. Sam Altman’s Ambitious Projections

• CEO Response and Bold Goals
  • Sam Altman claims OpenAI will surpass $20 billion in annualized revenue run rate by the end of 2025, and reach $100 billion by 2026.
    • “He said that by the end of this year, they're going to end it above $20 billion in annualized revenue run rate… by the end of the year, they'll be set to be making 20 billion, and in two years, they're gonna be set to making a hundred billion.” – Host [06:10]
  • Host notes such ‘run rate’ projections are highly speculative.

5. Compute Costs: The Billion-Dollar Question

• Inference and Training Costs:
  • OpenAI spent ~$3.8 billion on inference (the cost of running models) last year, up to $8.65 billion already in 2025.
  • Historically, OpenAI has relied almost exclusively on Azure for compute, but recent large-scale deals with CoreWeave, Oracle, AWS ($38 billion/7 years), and Google Cloud represent a diversification.
    • “They have one with Google Cloud, so they're making a lot more of these. But historically it's mostly been Microsoft Azure, just due to the nature of the $10 billion that Microsoft put in.” – Host [08:12]
• Cash vs. Credit Spend:
  • Most training uses non-cash 'credits' provided by Microsoft’s investment, while inference costs—needed for everyday AI use—require real cash outlays.
    • “Their training spend is tokens, but their inference, right, what they're using to service all of our requests … those are actually cash they got to pay.” – Host [09:10]

6. Implications for the AI Industry

• Potential Bubble and Sustainability Questions:
  • If OpenAI can't operate profitably—even at its scale—other firms may struggle, raising concerns about inflated valuations and a possible “AI bubble.”
    • “If OpenAI is really still running their models at a loss… this is like a massive deal for the investment world who is spending tons of money on these AI models.” – Host [10:15]

7. Official Silence

• No Comment from OpenAI or Microsoft:
  • Both companies declined to comment on the leaks or financial details.

Notable Quotes & Memorable Moments

• “In 2024 Microsoft received $493.8 million in revenue share payments from OpenAI…that number jumped to 865.8 million. And that's just in the first three quarters of this year.” – Host [01:10]
• “20% of the revenue OpenAI made, they were going to pay to Microsoft. …this has been reported and leaked by a lot of different people.” – Host [02:31]
• “So it's pretty hard to get a good estimate on how much they are currently kicking back and all of this.” – Host [04:02]
• “He said that by the end of this year, they're going to end it above $20 billion in annualized revenue run rate.” – Host [06:10]
• “Their training spend is tokens, but their inference…those are actually cash they got to pay.” – Host [09:10]
• “If OpenAI is really still running their models at a loss…this is like a massive deal for the investment world…” – Host [10:15]

Timeline of Important Segments

• [00:29] – Introduction to the leaked documents; detailing Microsoft’s revenue share with OpenAI
• [02:30] – Discussion of the 20% revenue share agreement and its origins
• [04:00] – Complexity of financial flows and difficulty in verifying “kickbacks”
• [05:05] – Analysis of revenue estimates and comparison of different sources
• [06:10] – Sam Altman’s projections and skepticism surrounding them
• [08:12] – Breakdown of compute costs; new deals beyond Microsoft Azure
• [09:10] – Clarifying the difference between 'credit' training spend and cash inference spend
• [10:15] – Reflections on the risk of an “AI bubble” and industry-wide implications

Conclusion

This episode unpacks the financial tightrope OpenAI walks as leaked documents show massive ongoing payments to Microsoft and eye-watering compute costs, raising new skepticism about AI profitability. As the host notes, “If OpenAI is not running at a profit, that means a lot of other people are not as well.” The revelations spark broader questions about the true economics of the AI boom, the potential for an industry-wide “bubble,” and the sustainability of current business models in artificial intelligence.