OpenAI projects that it will reduce the revenue share it pays to Microsoft from 20% this year to ~8% by 2030, potentially retaining over $50B more in revenue
The Information :
Context & Ripple Effects
OpenAI and Microsoft had already been negotiating the terms of their commercial relationship, including Microsoft's equity, OpenAI's IP, and the existing revenue share. OpenAI's reported effort to alter the exclusive cloud arrangement and future profit rights put the revenue split at the center of a broader renegotiation.
The projection gives a concrete economic target to that dispute: OpenAI wants more of the revenue generated by its products as it scales. Later reporting that capped Microsoft's payments at $38 billion suggests the two companies' financial ties were moving away from an uncapped participation in OpenAI's upside.
First-order effects
- OpenAI's model assumes a materially lower Microsoft take rate by 2030, potentially leaving it with more than $50 billion in additional cumulative revenue versus a continued 20% share.
- Microsoft faces lower prospective revenue participation from OpenAI, even as the companies negotiate the wider mix of ownership, technology access, and cloud terms.
Second-order effects
- A smaller revenue share increases the importance of other consideration Microsoft may seek in negotiations, such as equity, durable access to OpenAI technology, or cloud commitments.
- For OpenAI, retaining more top-line revenue would improve the economics available to fund its own growth and infrastructure obligations, reducing the share of future sales diverted to its strategic partner.
Third-order effects
- The case illustrates how frontier AI partnerships can evolve from simple distribution of revenue toward more bounded, renegotiated claims on a lab's future cash flows as that lab becomes commercially significant.
- If similar arrangements are repriced, AI infrastructure backers may favor clearer combinations of equity, capacity commitments, and capped payments rather than open-ended revenue shares; the balance will depend on who controls scarce compute and model access.
The trend: Frontier AI partnerships are being restructured as labs seek to convert early strategic dependencies into more sustainable economics and greater commercial autonomy.