Letter to AG Bonta and AG Jennings: Reject plan to give OpenAI nonprofit $100 billion stake in the for-profit
Attorney General Rob Bonta
Office of the Attorney General
California Department of Justice
Attn: Public Inquiry Unit
P.O. Box 944255
Sacramento, CA 94244-2550
Attorney General Kathy Jennings
Delaware Department of Justice
Carvel State Building
820 N. French St.
Wilmington, DE 19801
RE: OpenAI, Inc.
Dear Attorneys General Bonta and Jennings,
This letter is to follow up on our earlier correspondence about the corporate status of OpenAI.[1]
At least since November 2023, it has been clear that the combined OpenAI enterprises are functioning as a for-profit corporation, with OpenAI Nonprofit – which formally holds a controlling interest in the For-Profit – operating as little more than a rubber stamp of the for-profit. In our previous correspondence, we have detailed the voluminous evidence supporting this conclusion and urged that OpenAI Nonprofit should be dissolved, with the considerable value of its assets reinvested to charitable purpose.
In December 2024, OpenAI proposed that it would separate OpenAI For-Profit from OpenAI Nonprofit.[2] At that time, we warned that such an arrangement must involve the dissolution of OpenAI Nonprofit and payment under your supervision to independent charitable organizations.[3]
In May, OpenAI announced a reversal of its plans to separate OpenAI For-Profit from OpenAI Nonprofit, stating instead that OpenAI Nonprofit would continue to maintain a controlling interest in the for-profit affiliate. Importantly, the May announcement declared that OpenAI For-Profit would abandon its capped-profit structure, with a compensatory payment in shares of OpenAI For-Profit made to OpenAI Nonprofit.[4]
Yesterday, OpenAI announced that Open Nonprofit will be given a $100 billion stake in the for-profit.[5]
We are writing now to urge you to reject OpenAI’s plans categorically. They represent the continuation – and arguably a worsening — of the failed status quo.
If, however, you determine that the new arrangement is permissible, we urge you to require the payment from OpenAI For-Profit be made to a new, independent charitable entity or entities, rather than to OpenAI Nonprofit.
Under OpenAI’s current proposal, OpenAI Nonprofit would both maintain its non-shareholding controlling interest in OpenAI For-Profit and likely become OpenAI For-Profit’s largest shareholder. In this scenario, OpenAI For-Profit will be able to operate with a unique, double level of unaccountability. Because of the corporate structure and the shareholding interest, OpenAI For-Profit’s accountability would run to OpenAI Nonprofit and almost no one else. But as we have illustrated in our previous correspondence, and as we detail further below, despite the corporate formalities – and OpenAI Nonprofit’s purpose and legal duty – it is OpenAI For-Profit’s mission that runs OpenAI Nonprofit, not the other way around.
In other words, the new proposed arrangement is one that would both fail to honor OpenAI’s Nonprofit obligations and would insulate OpenAI For-Profit from the basic shareholder accountability attached to publicly traded firms.
OpenAI Continues to Evidence Its Prioritization of Profit and Abandonment of its Nonprofit Mission
Taken as a whole, OpenAI’s activities continue to evidence the complete and ongoing failure of OpenAI Nonprofit to uphold its mission and serve as anything more than a rubber stamp for OpenAI For-Profit. OpenAI’s actions make complete sense if OpenAI is understood to be a for-profit corporation prioritizing profit; they are completely incompatible with OpenAI’s claim that “the for-profit would be legally bound to pursue the Nonprofit’s mission, and carry out that mission by engaging in research, development, commercialization and other core operations,”[6] or with the notion that OpenAI Nonprofit is exerting control over the for-profit in the interest of advancing its guiding principles of safety and broad benefit. A few recent examples – building on the long list we have identified in earlier correspondence — follow.
First, OpenAI has consistently claimed that it recognizes the perils as well as enormous promise of artificial intelligence and that it understands and supports the need for appropriate guardrails. In 2023, for example, Sam Altman, the enterprise’s co-founder, testified before the U.S. Senate that: “OpenAI believes that regulation of AI is essential, and we’re eager to help policymakers as they determine how to facilitate regulation that balances incentivizing safety while ensuring that people are able to access the technology’s benefits.”[7] That sentiment arguably already indicated a compromise of the nonprofit’s mission, but it was widely interpreted to mean that OpenAI was different than Big Tech – precisely because of its nonprofit mission – and supportive of government regulation of AI technologies.
By the start of 2025, the company had reversed course entirely, instead advocating for a system of self-regulation and absolute preemption of state regulatory protections, including those that may be sought by California and Delaware. Astonishingly, OpenAI proposed a partnership with the federal government in which the federal government would impose absolute preemption on states in exchange for receiving “learnings and access” to information from AI companies. In this deal, state regulation would be completely preempted and the federal government’s role would effectively be limited to promoting AI development and safeguarding national security.[8]
Of course, as either a for-profit or nonprofit, OpenAI is free to advocate any views it chooses, on any topic. But OpenAI’s anti-regulatory views – including its effort to win federal policy that would eliminate the regulatory authority of California and Delaware — cannot be squared with OpenAI Nonprofit’s mission to “engage in research activities that advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return.”[9] Its 180-degree-turn to an anti-regulatory posture can only reasonably be understood as reflecting a commitment to the profit-seeking mission of the for-profit.
Second, OpenAI Nonprofit in January expanded its board, yet again underscoring that it perceives its primary mission as profit-making. The latest addition to the board is Adebayo “Bayo” Ogunlesi, a senior managing director at BlackRock. In welcoming him to the board, board chair Bret Taylor noted, “Bayo has an exceptional track record of driving success in organizations with global reach, expertly navigating complex and dynamic commercial landscapes.”[10] No doubt, Mr. Ogunlesi is highly accomplished; the point, however, is that he appears to have joined the nonprofit board to further a for-profit mission.
Third, the concerning multiple roles of Sam Altman – member of the nonprofit board, CEO of the for-profit affiliate, significant or primary investor in an unknown number of outside companies that may partner or intersect with OpenAI – continues to raise questions about how the nonprofit can abide by and uphold its nonprofit obligations. One evolving example is Sam Altman and OpenAI’s reported plans to help start a new brain-to-computer interface company, called Merge Labs.[11] The reporting on this is murky, because the deal has not been completed. What appears to be planned is that Altman will be a co-founder in his personal capacity and OpenAI’s venture arm will be a funder. Additionally, Merge Labs is also reportedly working with Alex Blania, who runs Tools for Humanity, the eye-scanning digital ID corporation that Sam Altman co-founded and currently chairs.
Leaving aside the important question about whether rushing into research on brain-to-computer technologies is compatible with OpenAI Nonprofit’s ethics commitments, this arrangement raises serious questions about how it comports with the legal duties of OpenAI Nonprofit and the obligations on Sam Altman as a member of the OpenAI Nonprofit board. Again, stipulating that details remain murky and are perhaps not fixed, what appears to be occurring is that:
- A for-profit affiliate of OpenAI, under the formal control of OpenAI Nonprofit, is investing in Merge Labs, a new, for-profit company.
- That new company, Merge Labs, is co-founded by Sam Altman, who likely has a profit stake in it.
- The new company, Merge Labs, appears poised to partner with Tools for Humanity, a company that Sam Altman co-founded and which he chairs.
- Sam Altman is simultaneously the CEO of OpenAI For-Profit and a member of the board of OpenAI Nonprofit.
Fourth, as we have noted in prior correspondence, OpenAI’s safety practices are out of step with the nonprofit mission’s prioritization of safety. Indeed, OpenAI is frequently the industry leader in introducing new, risky innovations to the market, forcing competitors to follow. One crucial concern is how chatbots interact with users, especially youth. Decisions to make chatbots seem human in their response (“anthropomorphization”) almost inevitably lead to user confusion and create the prospect of emotional dependence and manipulation.[12] As you have noted in recent correspondence with OpenAI, recent tragic events – a teen suicide and a murder-suicide that each followed extensive engagement with ChatGPT – underscore the problem and support the conclusion that “OpenAI and the industry at large are not where they need to be in ensuring safety in AI products’ development and deployment.”[13] Of course, as you note, AI safety is an issue for the entire industry. But what it special about OpenAI is that its nonprofit mission emphasizes safety; its failure to ensure sufficient safety standards is not just a moral and possibly legal failure, it is yet another indicator that the nonprofit mission has been subordinated to profit-seeking.
From these examples, and the long list of examples we have previously detailed — starting with the for-profit effectively firing the nonprofit board, after the nonprofit board had attempted to fire the CEO of the for-profit – one conclusion follows: OpenAI Nonprofit has subordinated its nonprofit mission to for-profit objectives. It should be dissolved and its assets – most significantly its controlling interest in OpenAI For-Profit — reallocated to independent charitable entities.
Payment Not to OpenAI Nonprofit
As we have argued previously, in a conversion that eliminates the capped profit structure, OpenAI Nonprofit should be compensated for this feature, over and above the control premium.[14] Now, OpenAI is proposing to maintain OpenAI Nonprofit’s controlling stake, but to eliminate the capped profit structure, and it agrees a substantial compensatory payment to the nonprofit must follow.
OpenAI proposes that the for-profit compensate the nonprofit $100 billion for sacrificing its claim to profits above the limits of the capped-profit structure. That is a substantial sum, but even this large amount requires scrutiny and may be insufficient. According to reporting, the sacrifice of the capped-profit structure was necessary for OpenAI to obtain new investment;[15] and OpenAI itself stated the move was designed to enable OpenAI For-Profit to gain access to hundreds of billions and perhaps trillions in financing.[16] The recent jump in OpenAI’s valuation – from $86 billion to an estimated $500 billion in less than two years – reflects the company’s new technologies, rapid gain of market share and future prospects. However, it also appears largely predicated on the change in OpenAI’s corporate structure. What would OpenAI For Profit’s value be today if the capped-profit structure were to remain in place? The delta between that number and $500 billion – which may very well be more than $100 billion – is the minimum of what OpenAI For-Profit should be required to pay for the elimination of the capped-profit structure.
Of even greater importance is that the payment from OpenAI For-Profit be made to an independent or multiple independent charitable entities, not to OpenAI Nonprofit. Given that it has proven itself subordinate to OpenAI For-Profit, a grant-giving OpenAI Nonprofit would effectively operate as an arm of OpenAI For-Profit, with a key if not primary purpose of promoting the interests of the for-profit. That would be an expanded perversion of the nonprofit form.
Indeed, it would be an inverse of the obligations of OpenAI Nonprofit: OpenAI Nonprofit is supposed to control the for-profit and it should serve the interests of the nonprofit. Instead, OpenAI now effectively proposes that the for-profit control the nonprofit and that it serve the interests of the for-profit.
The current proposal also would create profound corporate governance problems, It amounts to a complicated dual-class structure that would protect OpenAI For-Profit from external shareholder accountability.
OpenAI long ago abandoned its nonprofit purpose and mission, with the nonprofit subordinated to the for-profit. The best solution to the mutation of the nonprofit is to order it dissolved with its assets – both the equity interest from sacrificing the capped-profit structure and its controlling interest in the for-profit – allocated to new charitable entities that will advance the nonprofit’s original mission.
If you do not decide to require the dissolution of OpenAI Nonprofit, we urge that you require the major payment forthcoming from OpenAI For-Profit be made to one or more new, independent charitable entities, not to OpenAI Nonprofit. The logic for such a requirement is similar to the case for dissolution; allocating the payment to independent charitable entities would prevent the scenario of OpenAI Nonprofit working as an effective corporate foundation designed to advance the interests of the for-profit.
Thank you for your work preserving the interests of the charitable sector.
Sincerely,
Robert Weissman,
Co-President
[1] Public Citizen letter to Attorney General Rob Bonta, January 9, 2024, https://www.citizen.org/article/letter-to-california-attorney-general-on-openais-nonprofit-status; Public Citizen letter to Attorney General Rob Bonta, March 5, 2024, https://www.citizen.org/article/second-letter-california-attorney-general-openai-nonprofit-status-musk-lawsuit; Public Citizen letter to Attorney General Rob Bonta, June 6, 2024, https://www.citizen.org/article/june-2024-california-ag-openai-nonprofit-status-letter; Public Citizen letter to Attorney General Rob Bonta, September 30, 2024, https://www.citizen.org/article/fourth-letter-to-california-attorney-general-rob-bonta-openai-is-no-longer-operating-as-a-nonprofit; Public Citizen letter to Attorneys General Rob Bonta and Kathy Jennings, December 17, 2024, www.citizen.org/article/openais-for-profit-transformation-payments-must-go-to-new-independent-charitable-enterprise.
[2] OpenAI, “Why OpenAI’s Structure Must Evolve to Advance Our Mission,” December 27, 2024. https://openai.com/index/why-our-structure-must-evolve-to-advance-our-mission/
[3] Public Citizen letter to Attorneys General Rob Bonta and Kathy Jennings, December 17, 2024, www.citizen.org/article/openais-for-profit-transformation-payments-must-go-to-new-independent-charitable-enterprise.
[4] OpenAI, “Evolving Our Structure,” May 5, 2025, https://openai.com/index/evolving-our-structure/.
[5] Bret Taylor, “Statement on OpenAI’s Nonprofit and PBC,” September 11, 2025, https://openai.com/index/statement-on-openai-nonprofit-and-pbc.
[6] OpenAI, “Our Structure,” openai.com/our-structure.
[7] Testimony of Sam Altman before the Senate Committee on the Judiciary, May 16, 2023, https://www.judiciary.senate.gov/imo/media/doc/2023-05-16%20-%20Bio%20&%20Testimony%20-%20Altman.pdf
[8] Comment from Christopher Lehane to the Office of Science and Technology Policy, March 13, 2025, https://cdn.openai.com/global-affairs/ostp-rfi/ec680b75-d539-4653-b297-8bcf6e5f7686/openai-response-ostp-nsf-rfi-notice-request-for-information-on-the-development-of-an-artificial-intelligence-ai-action-plan.pdf.
[9] See Founding Document, OpenAI, Inc., https://rct.doj.ca.gov/Verification/Web/Search.aspx
[10] OpenAI, “Adebayo Ogunlesis Joins OpenAI’s Board of Directors,” January 14, 2025, https://openai.com/index/adebayo-ogunlesi-joins-openais-board-of-directors/
[11] Julie Bort, “Sam Altman, OpenAI Reportedly Back a Startup that Takes on Musk’s Neuralink,” TechCrunch, August 12, 2025, https://techcrunch.com/2025/08/12/sam-altman-openai-will-reportedly-back-a-startup-that-takes-on-musks-neuralink/.
[12] Rick Claypool, “Chatbots are not People: The Designed-in Dangers of Human-Like AI Systems,” Public Citizen, September 26, 2023, https://www.citizen.org/article/chatbots-are-not-people-dangerous-human-like-anthropomorphic-ai-report.
[13] Attorneys General Rob Bonta and Kathy Jennings, letter to OpenAI, September 5, 2025, https://oag.ca.gov/system/files/attachments/press-docs/2025-09-05%20-%20Letter%20from%20DE%20AG%20and%20CA%20AG%20-%20FINAL%20with%20NAAG%20Letter.pdf.
[14] Public Citizen letter to Attorney General Rob Bonta, September 30, 2024, https://www.citizen.org/article/fourth-letter-to-california-attorney-general-rob-bonta-openai-is-no-longer-operating-as-a-nonprofit
[15] Steven Tweedie, Ana Altchek and Rob Price, “OpenAI, facing pressure, announces its nonprofit will stay in control after all,” Business Insider, May 5, 2025, https://www.businessinsider.com/openai-non-profit-control-public-benefit-corporation-sam-altman-letter-2025-5
[16] OpenAI, “Evolving Our Structure,” May 5, 2025, https://openai.com/index/evolving-our-structure.