Microsoft's 100B Deal with OpenAI Means Microsoft May Control Superintelligence

Microsoft’s $100 billion profit milestone for OpenAI to certify that “artificial general intelligence” has been achieved is less about defining AGI and more about locking in control of the upside—an arrangement that looks heavily favorable to Microsoft and signals OpenAI’s urgency to secure scaling resources.

Under the deal, OpenAI’s board will only certify AGI once OpenAI reaches $100 billion in profits. That shifts the meaning of “general intelligence” from a technical threshold to a financial one. The practical effect is that Microsoft can tie the moment of AGI recognition to a long, measurable runway of profitability—giving Microsoft leverage over when the relationship changes and how much value flows back to it.

The transcript frames this as a strategic bet by Microsoft CEO Satya Nadella: if OpenAI can be pushed to a $100 billion profit outcome, Microsoft’s relationship with OpenAI can be reshaped at a point when the company has already generated extraordinary returns. The discussion highlights how rare it is for firms to reach $100 billion in lifetime profits. Only a small list of companies—Amazon, Saudi Aramco, Berkshire Hathaway, Apple, and Microsoft itself—have crossed that kind of cumulative profit figure. Microsoft’s lifetime profits are estimated at roughly $700 billion, making the $100 billion benchmark feel like a threshold that only a handful of firms in history have managed.

That rarity matters because it raises the question of whether the $100 billion requirement is a realistic proxy for AGI—or an intentionally stretched definition designed to benefit Microsoft. The transcript suggests that Microsoft lawyers pushed hard for this framing, implying the legal language gives Microsoft a strong position. It also notes that earlier deal terms left profit-sharing obligations “nicely undefined,” with an estimate that around 75% of profits would flow to Microsoft until AGI is achieved. By forcing a concrete financial line in the sand, Microsoft reduces ambiguity and potentially delays any shift in value allocation.

The transcript further argues that timing likely played a role. Nadella is portrayed as wanting to close the agreement in 2024 rather than renegotiate later, driven by the rapid pace of OpenAI’s progress—and a desire to avoid a more complicated conversation once the technology and bargaining dynamics evolve.

Finally, the discussion warns that the $100 billion profit milestone could be difficult even if OpenAI reaches advanced intelligence. Scaling AI can require massive spending on chips, data centers, and ongoing innovation. If OpenAI’s burn rate stays high, it may struggle to generate the kind of profits that satisfy the milestone for a long time—or possibly never. In that scenario, the deal’s structure would still keep Microsoft positioned as the dominant partner while OpenAI remains dependent on the resources needed to keep building.

Overall, the $100 billion profit trigger reads as a Microsoft-favorable mechanism that turns AGI certification into a leverage point—one that may reflect OpenAI’s current need for Microsoft’s support to scale, not just a neutral definition of general intelligence.