Overview
-
Partnership Structure: Microsoft owns 27% of OpenAI after investing ~\$13.4 billion since 2019, with OpenAI restructuring to create a \$130 billion nonprofit while converting to a public benefit corporation below
-
Exclusivity Terms: OpenAI's leading models remain Azure-exclusive through 2030 (or until AGI), while open-source models and consumer devices can be distributed elsewhere; revenue sharing continues until 2032
-
Compute Commitments: OpenAI plans \$1.4 trillion in compute spending over 4-5 years despite ~\$13 billion in 2025 revenue, with Sam Altman confident revenue will grow steeply to support these investments
-
AGI Timeline: Disagreement on timing - Nadella sees it as distant while Altman sounds more bullish, though both agree an expert panel would determine if/when AGI is achieved
-
2026 Outlook: Altman expects AI agents to handle multi-day tasks, small scientific discoveries by AI systems, and major advances in robotics and new computing form factors
Takeaways
Brad Gerstner's conversation with Sam Altman and Satya Nadella reveals the extraordinary scale of ambition driving the AI revolution. What fascinates me is how casually they discuss trillion-dollar commitments and the creation of one of the world's largest nonprofits as a byproduct of corporate restructuring.
What caught my attention was Altman's confidence about revenue scaling despite the eye-watering compute commitments. His point that OpenAI has beaten every business plan they've shown Microsoft suggests they see demand curves others don't - particularly around AI doing actual scientific discovery by 2026.
If we had 10x more compute, I don't know if we'd have 10x more revenue, but I don't think it'd be that far
The casualness with which they discuss AI making "small scientific discoveries" next year should give everyone pause. We're watching the construction of infrastructure for intelligence itself - not just another software platform.