The bankruptcy of the cryptocurrency exchange FTX in November 2022 had immediate philanthropic consequences. The exchange’s founder, Sam Bankman-Fried, had been the most visible large donor associated with the effective altruism movement, and the FTX Foundation and its associated giving vehicles had committed several hundred million dollars to causes ranging from pandemic preparedness to AI safety to political advocacy. After the bankruptcy, much of that committed funding either failed to materialize or was clawed back by the bankruptcy estate as preferential transfers.
By mid-2023 the second-order effects on the recipient organizations were becoming clear. Several pandemic-prevention and AI-safety nonprofits had built operating budgets around expected FTX commitments. Some downsized; some closed. The Future Fund, the regranting vehicle FTX had created, ceased operations entirely.
The OpenAI episode
In November 2023, the board of the OpenAI nonprofit removed the chief executive officer of OpenAI’s for-profit subsidiary — an action permitted under the unusual capped-profit governance structure the organization had adopted in 2019. Within five days, after employee protest and external pressure from investors in the for-profit subsidiary, the board had reversed itself, the chief executive had returned, and several of the directors who had taken the original action had resigned.
The episode was widely covered as a story about artificial intelligence policy, corporate governance, or technology personalities. Within the nonprofit governance community, it was something more specific: the most visible recent test of whether a nonprofit board could exercise the fiduciary control that the form requires of it when the underlying organization had grown to a scale that made such control practically unfeasible.
The OpenAI structure — a 501(c)(3) public charity that controls a for-profit operating subsidiary, with the nonprofit’s charter purposes of safe AI development meant to constrain the subsidiary’s commercial activity — was unusual but not unprecedented. The Mozilla Foundation, the Wikimedia Foundation, and several university research operations had used variants of the form. None had been tested at OpenAI’s scale or under comparable financial pressure.
The wider question
Both episodes pointed at a question the sector had been asking with increasing frequency: whether existing nonprofit governance norms were adequate to the scale and complexity of the largest contemporary charitable enterprises. The traditional 501(c)(3) board model — volunteer directors meeting four or six times a year, exercising oversight of a chief executive and senior staff — had been designed for organizations several orders of magnitude smaller than the largest today.
The conversation about governance reform that followed in 2024 produced no consensus. Some reformers argued for more demanding director qualifications, mandatory term limits, and standardized board-member orientation. Others argued for unbundling the governance and operational roles entirely, with paid governance professionals replacing the volunteer board model at large institutions. Neither approach gained enough traction to alter the legal default in any state.
The effect on effective altruism
The effective altruism movement itself continued, in muted form, after the FTX collapse. Open Philanthropy, the longtime regranting vehicle most associated with the movement, distanced itself from the more public faces of the previous era and continued its grantmaking on a more conservative public profile. The "earning to give" career-pathway argument that had once been central to EA recruitment lost most of its rhetorical force after 2022. Whether the movement would persist as an identifiable subculture or fold back into the broader rationalist and AI-policy spaces remained, by year’s end, an open question.