The AI Podcast

Episode Summary: OpenAI’s Financial Stability Threatened by Microsoft Compute Prices

Date: November 18, 2025
Host: The AI Podcast

Overview

In this episode, The AI Podcast investigates the financial relationships and challenges between OpenAI and Microsoft, focusing particularly on compute and revenue-sharing costs. Using leaked documents and recent reporting, the host exposes the scale of OpenAI’s payments to Microsoft, discusses what these numbers reveal about OpenAI’s business model, and explores the risks posed by high compute prices—and what that could mean for the broader AI industry.

Key Discussion Points and Insights

1. Revelations from Leaked Financial Data

• Leaked Documents & Source:

  • Tech blogger Ed Zitron shared documents exposing OpenAI's payments to Microsoft.
  • 2024: OpenAI paid Microsoft $493.8M in revenue share.
  • 2025 (first three quarters): Payments to Microsoft jumped to $865.8M—projected to be over $1B for the whole year.
    • [00:43]

    "At something like, you know, $865 million, you can assume it's going to be over a billion dollars that they're going to be paying Microsoft for this year alone."

• Revenue Share Details:

  • The 20% revenue share was agreed as part of Microsoft’s investment (reported, not officially confirmed).
  • Microsoft invested $1B initially, and later $10B, with total investments at $13B.
    • [01:28]

    "So total I think Microsoft all in is at $13 billion. And with that, I think especially with that $10 billion that they gave OpenAI, they agreed to a 20% revenue share."

2. Breakdown of Revenue Sharing and Compute Partnerships

• Reciprocal Revenue Sharing:

  • Microsoft gives OpenAI ~20% revenue from Bing and Azure OpenAI services.
  • OpenAI’s ChatGPT powers Bing; Bing search is integrated into ChatGPT.
  • Complex and mutually beneficial financial arrangements, but numbers are hard to verify.
    • [02:34]

    "So it's been said that OpenAI is giving 20% of their revenue because of the big investment back to Microsoft. And then Microsoft is giving 20% of revenue from things like Azure OpenAI service..."

• Opaque Financial Reporting:

  • Microsoft doesn’t break out Bing/Azure OpenAI earnings specifically.
  • Leaked numbers suggest these are NET revenue shares, not GROSS—Microsoft deducts some internal payments before revenue-sharing.

3. Massive Compute Costs and Revenue Analysis

• Estimating OpenAI’s Revenue:

  • Based on revenue share, OpenAI’s revenue was at least $2.5B last year (2024) and $4.33B in the first three quarters of 2025.
  • Conflicting Reports:
    • Some put 2024 revenue at $4B and the first half of 2025 at $4.3B; all reports are in the same ballpark.

• CEO Sam Altman on Revenue Claims:

  • Altman contests lower estimates, projecting $20B annualized run rate by year-end, and $100B by 2026.
    • [05:19]

    "He also says that the company is going to hit a hundred billion dollars by 2026 in just two years. So... 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 or they'll be making 100 billion. I don't know if that's actually gonna happen, but this is what he says."

• Exorbitant Compute Spending:

  • 2024: Estimated $3.8B in inference costs; 2025 (first 9 months): $8.65B.
  • Inference is the cost to generate responses for users ("servicing all of our requests").
  • Previously, OpenAI almost exclusively relied on Microsoft Azure; now diversifying (AWS, Google Cloud, etc).

4. Cashflow and Non-Cash Expenditure Dynamics

• Training vs. Inference Costs:
  • Training spend is mostly “non-cash” (prepaid credits included in Microsoft investment).
  • Inference spend is largely real “cash out the door.”
    • [09:24]

    "…a lot of it is credits that they're paying, paying Microsoft ... most of their spend is, you know, kind of these like tokens or credits that Microsoft already gave them. But there is a lot of spend that is largely cash when it comes to inference."

5. Implications for the AI Bubble and Industry Viability

• Profitability in Question:

  • If OpenAI operates at a loss, it's a red flag for the entire AI funding and investment ecosystem.
  • Raises concerns about inflated valuations and “AI bubble.”
    • [11:00]

    "Because if OpenAI is really still running their models at a loss ... this means this is like a massive deal for the investment world who is spending tons of money on these AI models."

• Both OpenAI and Microsoft Declined to Comment:

  • The episode closes by noting the lack of comment from both companies and highlighting the uncertainty ahead.

Notable Quotes & Memorable Moments

• [01:28] "So total I think Microsoft all in is at $13 billion. And with that ... they agreed to a 20% revenue share."
• [05:19] "He also says that the company is going to hit a hundred billion dollars by 2026 in just two years. ... I don't know if that's actually gonna happen, but this is what he says."
• [09:24] "...most of their spend is, you know, kind of these like tokens or credits that Microsoft already gave them. But there is a lot of spend that is largely cash when it comes to inference."
• [11:00] "If OpenAI is really still running their models at a loss, that this means this is like a massive deal for the investment world who is spending tons of money on these AI models."

Timestamps for Important Segments

• 00:20 – Leak overview and scale of OpenAI's payments to Microsoft
• 01:30 – Background on Microsoft’s $13B total investment and the origin of 20% revenue share
• 02:35 – Details on reciprocal revenue from Bing and Azure; discussion on opaque numbers
• 04:10 – OpenAI’s estimated revenue for 2024 and 2025; industry reporting compared
• 05:17 – CEO Sam Altman’s ambitious projections for OpenAI
• 07:06 – Compute cost breakdown (inference, training, partnerships)
• 09:24 – Nature of OpenAI’s cash vs. non-cash (credit) spend
• 11:00 – Broader implications for the AI bubble and industry viability

Podcast Tone & Language

The host maintains an informative, slightly skeptical tone—grounded in analysis, industry reporting, and a critical look at financial transparency. There’s a mix of accessible explanations for complex financial mechanics and a “pull back the curtain” mood when discussing leaks and industry secrets.

Takeaways

• OpenAI’s spending on Microsoft compute has grown rapidly, with current costs likely exceeding $1B/year.
• Microsoft’s investment (notably $13B in total) is recouped in part through a substantial 20% revenue share.
• While OpenAI's revenue and growth prospects are eye-popping, its sizable compute and inference costs—and the predominance of cash outflows for these—cast doubt on near-term profitability.
• The situation raises serious questions for investors and the perceived stability of the AI sector.