MartechARTICLE

OpenAI Is a Foundation-Controlled PBC Since 2025: What Changes for API Users

OpenAI Is a Foundation-Controlled PBC Since 2025: What Changes for API Users
Image: Eduardo Nogueira
Article · Martech
Reading
6 min
Tables
1
In short

Since October 2025, OpenAI Foundation controls the entire board of OpenAI Group PBC and holds 26% of its equity, valued at $130 billion. For anyone building on the API, this redefines who decides pricing, priorities, and safety limits.

What actually changed

OpenAI was founded in 2015 as a nonprofit organization, with the stated mission of ensuring that artificial general intelligence benefits all of humanity. In 2019 it created a for-profit subsidiary to scale research and product, but that subsidiary was always controlled by the nonprofit entity. On October 28, 2025, the company announced a recapitalization that formalizes that relationship in a different way: the nonprofit entity became OpenAI Foundation, and the commercial arm became OpenAI Group PBC, a public benefit corporation.

A PBC is required by charter to pursue its stated mission and consider the interests of all stakeholders, not just shareholder returns. According to OpenAI itself, the recapitalization was finalized after nearly a year of dialogue with the offices of the attorneys general of California and Delaware. That matters: the change didn't arise from market or customer pressure, it arose from regulatory scrutiny over converting a philanthropic entity into a business worth hundreds of billions of dollars.

Who's in charge: a permanent board the market doesn't choose

The power engineering is the central point for anyone who depends on the API. OpenAI Foundation holds special voting and governance rights that allow it to appoint every member of the OpenAI Group board and replace them at any time. Today that Foundation board includes Bret Taylor (chair), Adam D'Angelo, Paul Christiano, Sue Desmond-Hellmann, Zico Kolter, retired General Paul M. Nakasone, Adebayo Ogunlesi, Nicole Seligman, David Vélez, Robin Vince, and CEO Sam Altman.

To avoid what the company itself calls "fragmented governance," every Foundation director will also sit on the Group board, with one exception: Kolter serves exclusively on the Foundation, as a non-voting observer on the Group board. Within a year, a second Foundation director is expected to follow the same path, moving into an observer role. In practice, final control over the company that sells the API never passes through a shareholder, a stock exchange, or a market dispute: it resides in a self-perpetuating nonprofit board.

The safety committee that answers to no customer

The Safety and Security Committee (SSC) remains a committee of OpenAI Foundation, with governance authority over the safety practices of the entire OpenAI organization, including the Group. Zico Kolter, who chairs this committee, serves only on the Foundation. This means decisions about releasing or holding back model capability, applying stricter guardrails, or restricting use cases can come from a body that answers to no one paying the API bill and to no investor at all.

For an ordinary company, this kind of decision is normally pressured by customer churn or quarterly results. In OpenAI's structure, the pressure channel is different: the SSC answers to the mission, not to recurring revenue.

The math: 26% equity and a 15-year warrant

Upon closing the recapitalization, OpenAI Foundation came to hold a 26% stake in OpenAI Group, valued at roughly $130 billion based on the company's current valuation. That stake is in conventional common stock, which grows proportionally with the company's value, just like any other shareholder's.

In addition, the Foundation holds a warrant: if the Group's share price multiplies by more than ten times after 15 years, it receives a significant additional slice of equity. The design positions the Foundation to be the largest long-term beneficiary of OpenAI's success, which aligns financial incentives, but doesn't change the voting power at all, which is already permanent regardless of share value.

Before (for-profit subsidiary, 2019)After (2025 recapitalization)
Board controlNonprofit controls via governance structureFoundation appoints and replaces the entire Group board
Equity stakeCapped-profitConventional common stock, 26% + warrant
Safety committeeTied to the nonprofit structureRemains exclusive to the Foundation

What this means for those building on top of the API

Dependency risk changes shape, it doesn't disappear

The classic fear for anyone building on a platform is that the company gets sold, pivots, or disappears. That specific risk goes down: with the Foundation holding permanent and irrevocable voting power over the board, there's no scenario of a hostile sale or corporate dispute that removes OpenAI Group from the control of those who currently define the mission. That provides a structural predictability that purely private startups don't offer.

Pricing and roadmap move off the market's radar

In exchange, a less familiar risk enters the picture: product priority and access policy can be decided by a board whose legal mandate is to balance mission and the interests of all stakeholders, not to maximize API customer retention. Pricing remains subject to compute cost and competition, but decisions about what stays available, to whom, and under which guardrail go through a mission filter that no commercial contract can change. For founders betting their entire stack on a single vendor, that's a risk factor as relevant as SLA or price per token.

The counterpoint: a PBC is no guarantee of behavior

Delaware case law on public benefit corporations is still thin: courts rarely intervene unless a shareholder proves a clear violation of the duty to balance interests, and that's difficult to litigate in practice. OpenAI Group continues to seek capital, scale, and record valuation, competing directly with Anthropic and Google for market share, nothing in the legal structure prevents aggressive commercial behavior.

Worth remembering: the previous capped-profit model was also sold as mission protection and didn't stop OpenAI from raising tens of billions of dollars or from competing normally for customers. The more honest reading is to see the conversion to a PBC as a response to scrutiny from the attorneys general of California and Delaware, not as a spontaneous change in behavior toward whoever pays for the API.

What to watch going forward

For anyone deciding AI architecture and budget at a company, the governance structure has become a data point as relevant as uptime SLA. Worth tracking:

  • Changes in the composition of OpenAI Foundation's board, since it appoints and replaces the entire commercial board at any time;
  • Public decisions from the Safety and Security Committee about releasing or restricting model capability;
  • Signs that pricing or usage limits are being justified with mission language rather than market demand;
  • Keeping a plan B model (another vendor or an open model) for critical parts of the product, given that the governance lock reduces the risk of a hostile sale but not the risk of unilateral roadmap decisions.

The recapitalization doesn't by itself make OpenAI more or less trustworthy as a vendor. It just makes it clearer that, in this relationship, who decides the product's future isn't the market or the API customer: it's a nonprofit board with voting power that no funding round, no churn, and no acquisition offer can change.

Translated from the Brazilian Portuguese original · Read the original