After backlash, OpenAI has abandoned plans to restructure to remove control by its nonprofit entity. ProMarket reviews the history of OpenAI’s internal tensions to pursue profits over its founding purpose, artificial intelligence for the benefit of humanity, and what questions remain after the firm’s retreat.
On Monday, OpenAI announced that it had abandoned controversial plans to shed control by its non-profit parent entity. It will still convert its for-profit arm into a public benefit corporation (PBC) as intended. While a PBC signals that the firm will pursue goals beyond profit-maximization, how the firm pursues these goals and the metrics determining if the firm achieves these goals are at the discretion of company management.
OpenAI was originally founded as a nonprofit in 2015 with “the goal of building safe and beneficial artificial general intelligence for the benefit of humanity.” The company raised millions of dollars in donations, including from co-founder Elon Musk, and in credits and discounts from Amazon, Google, and Microsoft. For its first few years, OpenAI operated on donations.
OpenAI, which released its popular chatbot ChatGPT in 2022, has spearheaded the development of generative artificial intelligence. Generative AI uses large datasets to generate patterns that it can then reiterate in text, images, or video. In 2019, OpenAI CEO Sam Altman announced that the firm’s ambitions had exceeded what donations could support and that OpenAI would create a for-profit arm to attract investments. The for-profit entity would be completely controlled by its non-profit parent. It would also be under no obligation to pursue profits or make distributions to its investors. Any distributions it did make would be capped at 100 times investment. According to OpenAI, in 2019, “the for-profit [entity] raised an initial round of over $100 [million], followed by $1 billion from Microsoft.”
Purpose vs profits
In a recent Capitalisn’t episode and companion piece on ProMarket, Rose Chan Loui discussed how this change to OpenAI’s structure in 2019 commenced an ongoing internal struggle between the firm’s desire to commercialize and pursue profits to attract investment on the one hand, and safeguard its nonprofit founding purpose to ensure that the development of AI benefited humanity on the other. Chan Loui is the founding executive director for the Lowell Milken Center on Philanthropy and Nonprofits at the University of California, Los Angeles.
In 2020, OpenAI made its first product, GPT-3, a language learning model, available to the public. In 2022 came the first iterations of ChatGPT and DALL-E, an image generator.
The release of leading generative AI products began to pay dividends for OpenAI in 2023. Several hedge funds invested about $300 million into the company that spring. Microsoft promised $10 billion. However, Altman’s pursuit to commercialize OpenAI’s products and attract further investment worried some OpenAI employees that the product rollout was rushed and jeopardized the technology’s safety.
Tensions between purpose and profits erupted in November 2023 when OpenAI’s board of directors, after resignations from several of its members, removed Altman from the firm. In addition to disregarding safety concerns, the board claimed that Altman had failed to communicate adequately with them about product development and been abusive toward employees.
However, the board reinstated Altman days later after pressure from OpenAI’s investors and employees, the latter who threatened mass resignation. Those board members then resigned.
Altman’s return did not end the debate over OpenAI’s direction. In spring 2024, Musk sued Altman and OpenAI for a breach of contract, alleging that the firm had violated its purpose to pursue public benefit over profits. Experts doubted that Musk’s lawsuit would hold in court due to standing and legal vagueness in OpenAI’s founding commitments, but not necessarily because he was wrong. Musk initially dropped the lawsuit, only to revive it several months later.
The turmoil around Altman’s removal and Musk’s lawsuit brought to the fore questions about whether a nonprofit can abandon its purpose. ProMarket’s faculty director, Luigi Zingales, wrote in ProMarket after Musk’s lawsuit that such an allowance would let firms take unfair advantage of tax laws, from which OpenAI benefited as a nonprofit, and betray the expectations of donors, who gave OpenAI millions of dollars in initial funding to prioritize AI’s social benefits. Presciently, Harvard law professor Roberto Tallarita asked what corporate governance mechanisms can actually keep a firm from straying from its pronounced social purpose, and found the answer to be none.
In December 2024, OpenAI announced that it would restructure once more. The nonprofit arm would no longer have 100% control over the for-profit entity. Instead, it would have a minority stake. Meanwhile, the for-profit entity would convert into a PBC. According to OpenAI, the restructuring would open it to billions of dollars in investment to create a more sustainable firm.
Backlash and reversal
According to Chan Loui, the question surrounding its proposed restructuring was if OpenAI’s purpose to develop generative AI “for the benefit of all humanity […] can survive the enormous incentives and pressure to allow the company, now valued at $300 billion, to maximize profits without constraint, enriching its employees and investors, while casting aside the control of its nonprofit parent.”
In an open letter published in April, a group of former OpenAI employees, civil leaders, and academics, including Zingales, answered, “no.” The letter admonished OpenAI’s plans to restructure and was addressed to the attorneys general of California, where OpenAI is based, and Delaware, where it is legally incorporated. The letter warned the attorneys general that OpenAI’s plan to restructure threatened to violate its founding purpose, that OpenAI had failed to explain how its restructuring would advance its founding purpose, and that its proposed restructuring would remove corporate safeguards assuring OpenAI would pursue its original prosocial purpose.
On May 5, OpenAI announced that it would abandon plans to restructure so as to remove the nonprofit entity’s controlling status. Instead, the for-profit entity would still become a PBC but remain under the nonprofit’s control. In its explanation, OpenAI wrote, “We made the decision for the nonprofit to retain control of OpenAI after hearing from civic leaders and engaging in constructive dialogue with the offices of the Attorney General of Delaware and the Attorney General of California.”
Now what?
Questions remain about what OpenAI’s about-face will entail. The issue of attracting investment still carries as much weight as it did before. According to the New York Times, a recent investment from SoftBank has an option for the Japanese investment holding company to reduce its $40 billion investment to $20 billion if OpenAI does not complete its planned restructuring by the end of the year. Altman told reporters that SoftBank remains committed to its full investment under the revised proposed restructuring.
However, Bloomberg reported that Microsoft, OpenAI’s most important backer, has not yet approved the new restructuring plan. The two firms are in negotiations.
“We will need to know more about the governance structure and implementation to determine if this new proposal truly preserves the nonprofit’s purpose,” Chan Loui told ProMarket after OpenAI announced its U-turn. “Since it sounds like the nonprofit will have a large (but not the majority) equity interest in the Delaware public benefit corporation (PBC), it will need outsize voting rights on issues implicating the company’s plans for and implementation of AI development.”
Chan Loui elaborates further that “we will also want to hear more about the governance process, such as: Who will choose the board members of the nonprofit and the PBC? How will initial board members be replaced? What process will be put in place to ensure that the nonprofit board members are fully informed of the PBC’s technology development work?
For those who criticized OpenAI’s proposed restructuring last December, the firm’s reversal is a victory for the integrity of corporate governance and tax law. However, the devil is in the details, and Altman has proven himself determined to commercialize OpenAI.
Author Disclosure: The author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
