As OpenAI and Anthropic prepare for initial public offerings, concerns have emerged over their unique governance structures, which place control in the hands of "mission" directors who are not accountable to investors and may prioritize broader social goals over shareholder profits.

Both companies employ governance models where mission directors, often owning little or no equity, steer corporate decisions based on a stated commitment to benefit humanity, sometimes at the expense of financial returns. This arrangement marks a departure from traditional corporate oversight, where CEOs and investors typically hold significant authority.

The precedent for this model is often traced back to Ben & Jerry’s, the ice-cream maker sold to Unilever in 2000 under terms that allowed independent mission directors to preserve the company’s social mission. These directors could not be removed by Unilever and appointed their own successors, creating a self-perpetuating governance body that operated with limited accountability to Unilever. While conflicts were usually resolved quietly over two decades, a public dispute erupted in 2021 when the mission directors blocked the renewal of a licensing agreement with an Israeli business, citing inconsistency with the brand’s values. The ensuing legal battles and shareholder activism led to significant financial losses for Unilever, estimated at around $10 billion in market value, and culminated in the company transferring Israeli market rights and eventually spinning off its ice cream operations in 2025 to prevent future disruptions.

OpenAI’s governance structure is even more potent than Ben & Jerry’s was. Founded as a nonprofit in 2015 with a mission to safely develop artificial intelligence for humanity’s benefit, OpenAI established a for-profit subsidiary controlled by the nonprofit’s directors to raise capital. In 2023, these directors abruptly fired then-CEO Sam Altman, citing safety concerns, a move that nearly wiped out investor value and triggered internal upheaval. The board reversed its decision after significant employee pushback and pressure from Microsoft, highlighting governance instability. Subsequently, many researchers left to form competing AI ventures. Critics argue that the mission directors’ actions, while justified as safety measures, may have ultimately undermined OpenAI’s safety goals.

In 2025, OpenAI sought regulatory permission to remove nonprofit control over its for-profit subsidiary to facilitate its IPO but was only allowed limited restructuring. The for-profit arm, now a Delaware public benefit corporation (PBC), remains under the nonprofit’s control, with a board required to weigh safety and security considerations above profit. This structure maintains investor vulnerability to directors who are unaccountable and able to subordinate financial returns to mission-related decisions.

Anthropic follows a somewhat different approach. Its for-profit public benefit corporation, Anthropic PBC, is controlled by a "purpose trust" that appoints a majority of its board members. Unlike OpenAI, Anthropic’s trust explicitly considers investor interests alongside its mission. Additionally, Anthropic includes a provision allowing a supermajority of stockholders to terminate the trust and replace its board appointees—offering investors a potential mechanism to mitigate governance risks, although the exact threshold for such action remains undisclosed.

While these governance models aim to balance innovation and ethical considerations in a field with profound societal impact, some experts caution that unaccountable mission directors may expose investors to significant risks, as demonstrated by past conflicts. Furthermore, questions persist about whether such directors are essential to ensure AI benefits humanity, given that other tech firms and governments also influence AI development and regulatory frameworks.

Investors evaluating OpenAI and Anthropic ahead of their public offerings are advised to carefully examine these companies’ governance structures and consider the potential implications for investor protections and company decision-making priorities.