Titans Clashing: What OpenAI's Departure from Nonprofit Status Reveals About Real vs. Transactional Philanthropic Values

By Laurence A. Pagnoni, MPA Author, The Nonprofit Fundraising Solution | Former Faculty, NYU Heyman Center for Philanthropy, School of Professional Studies

Many nonprofit leaders heard that OpenAI was a nonprofit and felt a dissonance they could not quite name. These were our words — mission, public benefit, unconstrained by financial return — being used in a world that felt nothing like ours. A company now valued at hundreds of billions of dollars, at the center of a legal battle between two of the wealthiest people on earth, operating under a governance structure that was dissolving in real time. We looked away. The situation felt too distant, too compromised, too strange to engage with.

That feeling deserves a name and a framework — because what happened at OpenAI is not simply a story about artificial intelligence or Silicon Valley excess. It is a story about what the nonprofit form means, who uses it and why, and what is lost when it is treated as a corporate instrument rather than a public covenant. Nonprofit leaders have standing to name that loss, and this post tries to do so clearly. The place to begin is with what the nonprofit designation actually requires.

What Nonprofit Status Actually Means — and Why It Is Not Simply a Tax Category

The 501(c)(3) designation is not primarily a tax category. It is a seal of legitimacy conferred by the public through the federal government, and it carries obligations that distinguish it from other legal forms an organization can take. Mission primacy is required — the organization's assets and energy must be directed toward a charitable purpose, not toward the private benefit of its founders or investors. Private inurement is prohibited — no one in a position of control over the organization may use its assets for personal gain. And on dissolution, the assets must pass to another charitable organization, not to shareholders or founders. They are permanently locked to the public good.

That framework exists because the nonprofit form is built on public trust. The tax exemptions an organization receives, the deductibility of donations it accepts, the regulatory latitude it is granted — all of these rest on a social contract in which the organization promises to serve a mission rather than accumulate private wealth. When that promise is honored, the seal means something. When it is used instrumentally, it devalues the currency for organizations that have earned it.

The devaluation is not abstract. A donor who encounters one organization using the nonprofit form as a strategic convenience and another using it as a genuine expression of mission cannot easily tell the difference from the outside. The confusion erodes trust in the form itself — and that trust belongs to the broader sector, not to any single organization. 

There is an irony embedded in the name itself worth naming briefly. The "open" in OpenAI referred specifically to open-source research publication and open collaboration with other institutions — not to open governance, open accountability, or open democratic oversight. Elon Musk, who coined the name in December 2015 email exchanges with Greg Brockman, later said: "OpenAI was created as an open source (which is why I named it 'Open' AI), non-profit company. Not what I intended at all." The name was chosen for its simplicity and because the domain was available. It was not a philosophical claim.

Karl Popper's Open Society framework — the argument that no single person or institution should hold unchecked power over the conditions of collective life, and that institutions must be designed to be correctable when they are wrong — was nowhere in the founding conversation. By that standard, an organization developing artificial general intelligence under the governance of a small group of founders with minimal external accountability is precisely the kind of closed, uncorrectable institution Popper warned against, regardless of what it calls itself. The name "Open AI" has become, in retrospect, one of the more consequential misnomers in recent technology history — and part of what produced the confusion nonprofit leaders felt when they encountered it. We recognized our word. We did not recognize what was behind it.

Rose Chan Loui, founding executive director of the Lowell Milken Center on Philanthropy and Nonprofits at UCLA, put the legal point clearly: in the nonprofit world, purpose is a binding commitment made to the public, and the law does not make it easy to abandon it. It was precisely that principle that the OpenAI story tested.

The OpenAI Story: How a Nonprofit's Founding Promise Unraveled

OpenAI was founded in December 2015 as a nonprofit research institute, with a founding statement that promised artificial general intelligence development "unconstrained by a need to generate financial return." The founders, including Sam Altman and Elon Musk, wrote that because their research was free from financial obligations, they could better focus on a positive human impact. The founding intent, on the evidence available, appears genuine. Altman took no equity in the organization and accepted a salary of $76,000 per year as CEO — an unusual choice for a tech founder who was already a billionaire. He said publicly he did this because OpenAI was his childhood dream, and the record does not obviously contradict him.

What collapsed the nonprofit model was, in significant part, technological reality. Writing in the Financial Times, Simon Mundy used disclosed emails from the Musk v. Altman trial to trace the specific moment: in 2017, Google's GPU-cluster breakthrough made a single, large, hardware-intensive experiment the only viable path to frontier AI progress, rather than many smaller ones. OpenAI's nonprofit structure could not raise the capital to compete. By 2019, the organization introduced a "capped-profit" subsidiary to attract investment while claiming the nonprofit parent would retain control and that excess returns would flow back to the mission. By late 2024, it had restructured its core business into a for-profit benefit corporation, with the original nonprofit retaining approximately 20 percent of shares. Fortune reported that OpenAI changed its mission statement six times in nine years, finally removing all references to safety in its 2025 IRS Form 990 — the last time the company claimed tax-exempt status.

The nonprofit sector can extend reasonable understanding to the technological argument — compute costs are real, and the capital requirements of frontier AI research genuinely exceed what philanthropic funding can provide; but understanding is not absolution, and the full picture is more complicated than structural necessity alone.

The governance constraint that the nonprofit form imposed on Altman personally was that he had to preserve a majority disinterested board to maintain nonprofit status — which meant he could not hold equity in OpenAI. He acknowledged this himself on the All-In podcast in May 2024. The restructuring that was framed as technological necessity was also the restructuring that removed that constraint. Reuters reported that Altman would receive equity as part of the for-profit conversion. He subsequently told staff he would not receive a "giant equity stake" — carefully worded. The nonprofit sector is not required to adjudicate whether Altman's motives were pure or mixed. What we can note is that the outcome served his personal interests alongside the organization's competitive ones, and that the organization used the nonprofit form for nine years of reputational and regulatory benefit before discarding the governance obligations that came with it.

Luigi Zingales, faculty director at ProMarket, wrote that allowing such a conversion would let firms take unfair advantage of tax laws from which OpenAI benefited as a nonprofit, and would betray the expectations of donors who gave OpenAI millions of dollars in initial funding to prioritize AI's social benefits. 

Alnoor Ebrahim, a professor at Tufts University's Fletcher School who first noticed the mission statement changes, was more pointed: in his view, these changes explicitly signal that OpenAI is making its profits a higher priority than the safety of its products. 

A dissenting voice worth acknowledging comes from Cortney Nicolato, a Northeastern University lecturer, and CEO of the United Way of Rhode Island, who argued that OpenAI is still keeping its nonprofit legacy intact in many ways, noting the OpenAI Foundation's 26 percent equity stake and its $1 billion commitment to health care and AI resilience programs. That reading is more charitable — and it is worth holding alongside the more critical assessments rather than dismissing it.

That is the pattern that produces the dissonance nonprofit leaders felt. Not villainy, necessarily, but instrumentality — the nonprofit form as a tool picked up when useful and set down when it stopped serving its purpose, with legal scholars, economists, and nonprofit practitioners divided on whether what remained constitutes a genuine honoring of the original public commitment.

Elon Musk and OpenAI: What Transactional Philanthropy Looks Like in Practice

Elon Musk's role in the OpenAI story is worth examining separately because it is both simpler and more revealing than Altman's. He co-founded the organization, contributed approximately $38 million through his foundation on utilitarian principles, then departed in 2018 citing disagreements with the organization's direction. He subsequently sued OpenAI multiple times seeking to block or reverse the for-profit conversion — not, the record suggests, from principled attachment to the nonprofit form, but from competitive interest. He launched his own AI company, xAI, which directly competes with OpenAI, after leaving the organization. A lawsuit seeking to reverse a competitor's governance structure is a different instrument than a principled defense of nonprofit law.

Musk's posture toward philanthropy more broadly clarifies what transactional values look like in practice. When a viral post noted MacKenzie Scott's $26.3 billion in donations had made her one of the biggest individual donors in history, Musk responded agreeing with the characterization that she was making the world a worse place, saying: "Sadly, yes." He has argued that giving money away well is extraordinarily difficult and that it is easy to give for the appearance of goodness rather than the reality of it. The observation is not without some philosophical content — stewardship of philanthropic capital is genuinely challenging. But it sits uneasily alongside his own foundation's record, his lawsuit behavior, and his public dismissal of one of the more substantively documented philanthropic programs in contemporary American history.

Transactional values in philanthropy are not always cynical at the outset. They can begin as genuine conviction and gradually become something else as circumstances change and interests diverge. What distinguishes them from real philanthropic values is the test they fail when the form stops serving the founder's purpose: the transactional philanthropist sets the form down. The principled one asks what obligations it carries. That distinction is what makes MacKenzie Scott worth examining at some length.

MacKenzie Scott and the Case for Trust-Based Philanthropy

MacKenzie Scott has given away $26 billion since 2019 — $7.2 billion in 2025 alone. Scott requires no follow-up from recipients. There are zero buildings, plazas, or endowed chairs named after her. She has declined to speak with the media beyond the essays she writes each time she announces a new round of grants. A defining feature of her philanthropy is its lack of restrictions — her donations typically come without conditions, allowing recipient organizations to determine how best to use the funds. Noni Ramos, CEO of Housing Trust Silicon Valley, described the experience this way after receiving a $30 million gift: "Unlike traditional funding processes that often involve lengthy applications, specific restrictions, and reporting requirements, her style empowers organizations like ours to determine how best to direct funds quickly and innovatively to address pressing issues."

Scott did not use the nonprofit form. She funded nonprofits, trusted them, and stepped back. Her approach reflects what real philanthropic values look like when held consistently: the mission of the recipient organization is primary, the donor's ego is absent, and the measure of success is what happens in communities rather than what appears on a donor's letterhead. As Inside Philanthropy has written, Scott's evolving apparatus remains something more unusual — a high-dollar philanthropic engine built around the proposition that giving up power may be the point.

The contrast with Musk is not ultimately about the scale of giving. It is about the underlying theory of what philanthropy is for. Musk's theory is utilitarian and control-oriented — he gives toward outcomes he has defined as beneficial and objects when the results diverge from his expectations. Scott's theory is relational and trust-based — she identifies organizations doing genuine work, provides capital without conditions, and trusts the people closest to the problems to deploy it well. One of these theories is compatible with the nonprofit sector's foundational values. The other treats those values as instrumental.

Real Philanthropic Values vs. Transactional Values: A Framework Nonprofit Leaders Can Use

The OpenAI story is useful to the nonprofit sector not because it is typical — it is not — but because it makes visible a distinction that is usually harder to see. Real philanthropic values and transactional values often look similar from the outside, at least at the beginning. The framework for telling them apart over time has three tests.

The mission primacy test. Does the organization's behavior over time reflect genuine subordination of personal and institutional interests to the mission, or does the mission language track the interests of the founders and major stakeholders? At OpenAI, the mission language remained largely consistent even as the governance structure moved steadily in the direction of founder and investor benefit. Chan Loui noted that Nonprofit OpenAI's purpose was uniquely and very specifically tied to its control of For-Profit OpenAI's operations — making the restructuring not merely a change in activities but a fundamental shift in purpose. That gap between stated mission and structural behavior is the signal worth watching in any organization.

The governance test. Does the board exercise genuine independent oversight, or does governance exist to legitimate decisions already made? The 2023 ouster of Altman by the OpenAI board — and his reinstatement five days later after investor pressure — tested the nonprofit's governance in a way that few organizations experience publicly. The outcome suggested that the formal governance structure was not as independent as the nonprofit form requires, a concern Darryll K. Jones, a law professor at Florida A&M University College of Law, raised specifically in Bloomberg Law: when nonprofit fiduciaries are both seller and buyer in a conversion, the transaction must be managed to negate conflicts of interest.

The dissolution test. On dissolution, where do the assets go? For a genuine nonprofit, the answer is locked: to another charitable organization. For an organization that has converted to a for-profit benefit corporation while retaining a nominal nonprofit minority shareholder, the answer is considerably less clear. The test of real philanthropic values is what happens to the assets when the founding vision is no longer convenient.

These three tests are useful not only for assessing organizations like OpenAI but for any nonprofit leader conducting an honest self-assessment of their own organization. The seal of legitimacy is not self-conferring. It must be earned continuously, through governance that exercises real independence, through mission behavior that does not drift toward founder or funder convenience, and through structures that will honor the public covenant even when it is no longer easy to do so.

Why the Nonprofit Sector Has Standing to Reclaim What the Seal of Legitimacy Means

The nonprofit sector has standing to feel offended when the form is used instrumentally — not as a grievance, but as a signal that something worth protecting is being devalued. The offense is the appropriate response to what happens when decades of sector-wide trust-building is borrowed by organizations whose commitment to the form is conditional. When that offense leads to looking away, as it did for many of us with OpenAI, the sector loses an opportunity to name clearly what it stands for.

What the sector stands for is a public covenant — a promise to the communities we serve that the organization's assets and energy are permanently directed toward the mission, that no private interest supersedes it, and that the form will be honored even when it stops being convenient. That covenant is what the seal represents. It is worth defending, worth naming, and worth distinguishing from the uses to which it has recently been put.

The titans can clash all they want over artificial intelligence, competitive advantage, and the governance of a half-trillion-dollar enterprise. The nonprofit sector knows what the form means. It is time we said so.

What has this episode clarified for you about the nonprofit form and what it should mean — and how do you think the sector should respond? Share your experience in the comments section of the website.

A Note on Sources

The Financial Times article referenced throughout this post is: Simon Mundy, "How the Dream of a Nonprofit OpenAI Died," Financial Times, May 2026. The article draws on disclosed emails from the Musk v. Altman trial and is the most detailed account available of the specific technological moment — Google's 2017 GPU-cluster breakthrough — that made the nonprofit model for frontier AI development commercially untenable. 

Additional expert commentary cited: Rose Chan Loui, ProMarket, April 2025; Luigi Zingales, ProMarket, 2024; Alnoor Ebrahim, quoted in Fortune, February 2026; Darryll K. Jones, Bloomberg Law, November 2024; Cortney Nicolato, Northeastern University News, November 2025; Noni Ramos, quoted in Fortune, December 2025; Inside Philanthropy, MacKenzie Scott profile, June 2026.

This post is offered freely for educational purposes. Please share it with executive directors, board members, development staff, and colleagues who may find it useful — provided the author's byline remains intact: By Laurence A. Pagnoni, MPA. Reproduction in publications, training programs, or institutional materials requires attribution.