This is the first part of our three part series on covering the different aspects of the threats posed by Anthropic’s IPO and other mega IPOs.
Tech Oligarchs’ Bid for Control at Our Expense
Part 1: State Corporate Law
By Natalia Renta, Associate Director for Corporate Governance & Power
Anthropic is legally incorporated in Delaware, the chosen state of incorporation for a supermajority of companies that go public. In recent years, tech oligarchs have challenged Delaware’s status as the state of choice for public company incorporations, triggering a race to the bottom in state corporate law that Delaware has not been immune to. While Texas and Nevada are much worse for ordinary shareholders, Delaware has also changed its laws to strengthen the power of corporate insiders like Anthropic’s founders while decreasing the power of regular shareholders to provide input and hold management accountable for misconduct.
Recent changes to Delaware corporate law
As of 2024, Delaware allows companies incorporated in the state to enter into contracts granting prominent shareholders the right to make decisions historically reserved for boards of directors. The basics of corporate law are that shareholders elect boards of directors who in turn hire and oversee executives. These directors owe a fiduciary duty to all shareholders. Contracts that allow individual shareholders to unilaterally make important corporate decisions upend this structure. Delaware very likely made this change to their law at the behest of private equity firms to quickly overturn a decision by a Delaware judge that a shareholder could not contractually usurp a long list of critical decisions belonging to boards of directors. In practice, this means that Dario and Daniela Amodei and other AI founders could enter into contracts that give them the power to override their board’s decisions.
In 2025, following pressure from tech oligarchs, Delaware embarked on a major overhaul of its corporate law. SB21, referred to by opponents as the “Billionaires’ Bill,” makes it much easier for controlling shareholders, directors, and executives to avoid scrutiny for self-serving actions. It also makes it much more difficult for regular shareholders to obtain internal communications such as emails and text messages, which can be critical for getting the necessary evidence to legally challenge wrongdoing by corporate insiders. This all boils down to regular shareholders and the public having limited visibility and say in the companies shaping the future of AI — while a few founders call the shots.
To make matters worse, the Regulatory Sandbox Subcommittee of the AI Commission created by the Delaware legislature has proposed creating an AI regulatory sandbox, so companies incorporated in Delaware, like Anthropic, could create another limited liability entity supposedly run by AI agents and ask an oversight committee to exempt it from certain Delaware regulations, laws, or licensing agreements. This would create a whole new layer of impunity for AI run amok, leaving people who are harmed by AI with limited or no redress.
Veneer of responsibility
Anthropic has an unusual corporate structure. It is incorporated as a public benefit corporation with the purpose of “responsible development and maintenance of advanced AI for the long-term benefit of humanity.” The company created a five-member trust composed of “experts in AI and ethics” with the ability to select and remove a majority of Anthropic’s board members over time. At first blush, this structure seems to confirm Anthropic’s public narrative that it is the responsible AI company that should be trusted because its founders are concerned about AI’s impacts in the world.
Upon closer examination, however, Anthropic’s corporate structure, combined with Delaware’s founder-friendly laws, add up to control being locked in the hands of a few — the founders and members of the trust — with very little opportunity for regular shareholders or the public to weigh in. This power structure might soon be made permanent: recent news reports covering Anthropic’s leaked IPO documents indicate that Anthropic’s founders are poised to have majority — 50.1% — voting control when the company goes public.
Wrestling control over AI from tech oligarchs
Delawareans are fighting back. In the recent primary election, a backlash unseated several of the Delaware legislators who championed changes to the state’s corporate law to the benefit of tech oligarchs. The winners made the case that their opponents put the interests of billionaires above working people in Delaware. The incoming class now has a chance to do the opposite. Recognizing the high degree of influence the state has over corporate law, policymakers should resist and reverse the race to the bottom. Delaware and other states should also regulate AI in the public interest and strengthen and deploy their corporate accountability tools.
At the federal level, Congress should legislate strong standards so AI is developed and deployed in the public interest, not according to the unchecked whims of founders and the financial interests of their billionaire financiers. A critical part of this effort needs to be to set a corporate governance floor for AI companies (and public companies more generally) so that regular shareholders, workers, and affected communities have the power to further their interests in decisions that will significantly affect their lives. This floor should include robust governance and litigation rights for regular shareholders at minimum, and could include additional requirements, such as union neutrality and a certain percentage of workers on corporate boards.
