OpenAI swaps its profit cap for a public benefit corporation structure

Since October 2025, OpenAI has gone from a capped-profit company controlled by a foundation to a public benefit corporation with conventional shares. The design holds direct lessons for anyone raising AI-native rounds in Brazil.

A restructuring from over a year ago that still sets the rules of the game

On October 28, 2025, OpenAI closed a recapitalization that changes the incentive logic of the entire company. According to OpenAI's official governance page (openai.com/our-structure), the nonprofit founded in 2015 became the OpenAI Foundation, and the commercial subsidiary created in 2019 became OpenAI Group PBC, a public benefit corporation.

The difference isn't just a name change. A PBC is required by statute to advance a stated mission and consider the interests of all stakeholders, not just maximize shareholder return. It's the same corporate form used by Anthropic and Patagonia, but at a much larger scale: the source cites the Foundation's stake in the new structure as worth roughly $130 billion.

The process wasn't quick. OpenAI itself describes the change as having been "completed after nearly a year of constructive dialogue with the offices of the attorneys general of California and Delaware," a detail that signals how closely regulators were watching an entity that promises public benefit while moving billions in venture capital.

What existed before, and why it no longer held up

Since 2019, OpenAI's commercial structure had operated under what's known as the capped-profit model: investors put money into a for-profit subsidiary, but their returns were capped at a multiple of invested capital, and final control rested with the nonprofit's board. It was a design meant to attract capital without giving up the original mission of safety in artificial general intelligence.

The problem is that this model doesn't scale well once the capital math climbs into the tens of billions of dollars. A return cap is a real friction point for institutional investors used to full participation in the upside, especially at a company competing for talent and capital against conventionally structured rivals.

Under the new structure, according to OpenAI, "all OpenAI Group equity holders now hold the same type of traditional equity, which shares proportionally in, and grows in value with, OpenAI Group's success." In other words: the cap is gone, and whoever invests in OpenAI Group gets full exposure to the company's growth, with no multiple limit.

How the balance of power ended up

The engineering behind the new structure tries to answer a hard question: how do you remove the profit cap without handing control of the company to people who only want a return? OpenAI's answer has three parts.

• OpenAI Foundation retains special voting and governance rights that allow it to appoint all members of OpenAI Group's board and replace them at any time.
• The Foundation holds a 26% conventional equity stake in OpenAI Group, valued at approximately $130 billion at the company's current valuation.
• The Foundation also holds a warrant: if OpenAI Group's per-share value grows more than tenfold over 15 years, it receives an additional significant tranche of equity.

This design makes the Foundation, in the company's own words, "the largest long-term beneficiary of OpenAI's success" among all stakeholders. It's a way of aligning the financial incentive of the entity that controls governance with the business's actual growth, rather than leaving it dependent only on donations or a fixed cap.

The Safety and Security Committee (SSC) remains a Foundation committee, not a Group one, and its chair, Zico Kolter, moves to a non-voting observer role on the commercial board. The stated goal is to avoid "fragmented governance": nearly all Foundation directors also sit on the Group board, with the exception of Kolter and, within a year, a second director who will move to an observer role.

The detail that matters to Brazil: who's at the table

Among OpenAI Foundation's independent directors is David Vélez, founder and CEO of Nubank (a Brazilian digital bank). His presence on the board that controls one of the most valuable AI companies in the world is concrete evidence that the conversation about generative AI governance isn't only happening in Silicon Valley: prominent Brazilian founders are already helping shape these rules from the inside.

This matters because the logic OpenAI has formalized, mission and commercial return advancing together, without one limiting the other through an artificial cap, is exactly the tension any Brazilian founder raising an AI-native round faces today: how to promise impact and responsible governance to attract certain investors, without scaring off those who want full participation in the upside.

The counterargument no official statement makes

The more critical reading of the change is simple: removing the profit cap benefits directly those who are already invested, including Sam Altman and the round's major backers, more than it benefits the public the original mission promised to protect. The capped-profit structure existed precisely to contain that kind of capture.

This critique gains weight when you look at the approval process: according to OpenAI itself, negotiations with the attorneys general of California and Delaware took nearly a year before closing, a sign that the structural change required real regulatory scrutiny, not just a public-relations announcement.

OpenAI's response to that critique lies in the mechanisms themselves: special voting rights that don't disappear with equity dilution, and a warrant that lets the mission entity grow alongside the company instead of being stuck under a cap. Whether that's enough of a guarantee that the mission won't be subordinated to financial return is something only time (and the next funding round) will show.

What this teaches anyone raising money right now

For a Brazilian AI-native startup founder negotiating a round with an international fund, the OpenAI case offers three lessons in corporate design that go beyond the hype.

1. A return cap is capital friction: if the business thesis depends on attracting large checks and sought-after talent, an artificial multiple cap can cost more in opportunity than it seems to protect in mission.
2. Special voting rights survive dilution: a foundation or holding company with the power to appoint and replace board members protects control even as its equity stake falls proportionally with new rounds.
3. Aligning the financial incentive of whoever controls governance with the business's growth (through real equity, not just donations or a cap) tends to last longer than relying on the founder's good will.

None of these lessons resolves the underlying tension between a stated mission and the pressure for returns. But they show that this tension can be designed into contracts and board clauses, instead of being left to personal trust in the founder, which is exactly what OpenAI tried to do after nearly ten years of an improvised structure.