Anthropic Plans Dual Mechanism to Lock Control and Resist External Shareholder Pressure Post-IPO
Anthropic is reportedly preparing to issue high-vote shares to CEO Dario Amodei and other co-founders to resist external shareholder pressure after a potential IPO.
The arrangement has not been finalized, and details such as the multiple of voting rights per share, the distribution among founders, and whether a sunset clause will be set have not been disclosed; the company may go public as early as the end of September, but Anthropic has not publicly confirmed the timeline for the IPO.
Reports indicate that Amodei currently holds about 2% of Anthropic's shares, with external financing diluting the economic holdings of the founding team. High-vote shares would allow him to retain greater voting influence without holding corresponding economic rights. reuters+1
Anthropic also plans to maintain a Long-Term Benefit Trust. This independent entity does not hold economic rights but nominates a majority of board members through special Class T shares; the company confirmed in April that trust-nominated directors hold a majority on the board.
Anthropic operates as a Public Benefit Corporation, legally required to balance commercial success with public interest. The combination of high-vote founder shares and trust board control means that ordinary public shareholders, even if they provide significant equity capital, may not gain governance influence commensurate with their economic rights.
From a market mechanism perspective, the dual control structure benefits management and the trust in maintaining long-term AI safety investments, rejecting short-term profit-driven commercialization demands, and reducing the feasibility of hostile takeovers; the trade-off is a weakening of public shareholders' voting rights, governance correction capabilities, and board accountability. Beneficiaries include the founding team, the trust, and strategic investors who prefer a long-term R&D path; those under pressure are public market investors demanding one share one vote, board independence, and capital allocation transparency.
Source: Public Information
ABAB AI Insight
Anthropic was founded in 2021 by Dario Amodei and other former core members of OpenAI, subsequently receiving strategic funding support from Google, Amazon, and others, and entering the enterprise-level large model market with the Claude series models. Unlike typical tech startups, Anthropic embedded "charter-style governance" into its corporate structure from the outset: the company is set up as a Public Benefit Corporation and established a Long-Term Benefit Trust to oversee its public interest goals. The 2023 incident where the OpenAI board ousted Sam Altman and then quickly restored him under pressure from employees and investors highlights that when governance rights, commercial capital, and talent control are misaligned in cutting-edge AI companies, safety mission arrangements can be rapidly rewritten by operational realities.
In terms of capital pathways, Amazon and Google not only provide funding to Anthropic but also offer cloud computing and model distribution infrastructure; the company converts its model training investments into recurring revenue by selling the Claude API and subscription services to enterprises. The founders' approximately 2% stake indicates that the traditional "founder-majority control" pathway has failed, and high-vote shares attempt to separate control from economic capital; the Long-Term Benefit Trust further ensures that board appointment rights are not entirely influenced by founders, cloud vendors, or future public shareholders.
Historically, Meta has used dual-class shares to allow Mark Zuckerberg to maintain majority voting power with minority economic rights, while Alphabet has maintained the influence of founders and early management through multiple classes of stock; however, Anthropic's Long-Term Benefit Trust adds an extra layer of control over the board by a "non-economic rights entity" compared to Meta and Alphabet. At this stage in the industry, cutting-edge model companies are transitioning from private financing and strategic support from cloud vendors to a phase that requires public market funding to support computational expansion; Anthropic is attempting to lock in its governance structure before scaling up financing to prevent short-term performance pressures from reshaping model releases and safety investments post-IPO.
Essentially, this is about capital concentration. Training cutting-edge models requires centralized computing power, data centers, power contracts, and massive capital, and financing demands will quickly transfer economic rights to cloud giants, private investors, and public markets; if control rights also shift, safety and product direction will be dominated by marginal capital returns. By using super-voting rights and trust board powers to sever the traditional link between "capital contribution—decision-making power," Anthropic can protect its long-term path but will make it difficult for capital to impose constraints on management through voting.
ABAB News · Cognitive Laws
Capital can buy returns, but not necessarily direction
When capital demands speed, control determines the braking power
Equity distribution allocates wealth, voting rights allocate the future