Anthropic CEO Dario Amodei looks on after a meeting with French President Emmanuel Macron during the AI Impact Summit in New Delhi on February 19, 2026.
Ludovic MARIN/AFP via Getty Images
Anthropic is preparing to give CEO Dario Amodei and its six other co-founders a new class of stock carrying enhanced voting rights ahead of an IPO expected as soon as late September — and anyone buying shares in what could be the largest public market debut in history should understand, before allocating a dollar, that they will own a financial instrument with almost no ability to influence how the company is run. Reuters and Bloomberg independently confirmed the plan in separate reports published August 18.
The super-voting plan, first reported by The Information on August 18 and confirmed independently by Bloomberg, represents the first time Anthropic’s leadership would hold enhanced voting power. It arrives at a specific moment: the company’s Long-Term Benefit Trust, the independent body that elects a majority of Anthropic’s board of directors, has just shrunk from four trustees to three following a trustee’s move to the company’s own executive team. Anthropic’s company page confirms three trustees: Neil Buddy Shah, Richard Fontaine, and Ben Bernanke. The combined effect — founders controlling votes on one axis, a three-person trust controlling board composition on another — gives public shareholders effectively no lever to pull if Anthropic’s direction disappoints.
What Founder Super-Votes Mean at a Company With 2% Founder Equity
The case for super-voting shares at Anthropic is arithmetically unusual. CEO Dario Amodei owns roughly 2 percent of the company, according to a source cited by The Information — a stake that years of dilutive fundraising reduced from what would have been a normal founder position. The company has pulled in more than $85 billion in venture capital since its 2021 founding, including a round that closed at $65 billion Series H in May 2026 at a post-money valuation of $965 billion. At 2% equity, a standard one-share-one-vote structure would leave Amodei without the voting power needed to withstand pressure from outside shareholders once the company goes public.
The dual-class structure being prepared would solve that problem by design. Such structures are not unusual in technology companies: at Meta, Zuckerberg holds 60 percent voting control through super-voting shares despite owning a fraction of the company’s economic interest. SpaceX’s structure gives Elon Musk approximately 85% of total voting power. The Council of Institutional Investors advocates against dual-class structures without time-based sunset provisions of seven years or less, arguing that they reduce management accountability to the economic owners of the business.
The specific vote ratio for each founder share has not been disclosed. Anthropic has not publicly commented on the plan, and The Information noted that the specifics “could not be learned” and the structure could still change.
Why the Trust’s Three-Member Headcount Matters Right Now
The Long-Term Benefit Trust (LTBT) is the element of Anthropic’s governance that makes it structurally unlike any company preparing to go public. The Trust holds a special class of stock — Class T shares — that carries no economic value but grants its trustees the authority to elect a growing number of Anthropic’s seven board members, ultimately reaching a majority. In plain terms: a body of three people who own no equity in Anthropic has more direct power over its board than either the founders or outside investors, once the Trust’s appointment authority is fully exercised. The Class T shares elect board majority under the milestone schedule established in the company’s original charter.
When Anthropic announced the LTBT in September 2023, it had five trustees. Over time, several members cycled through the body, including Jason Matheny, Paul Christiano, and others. The Trust stood at four members as recently as July 9, 2026, when Bernanke joined the Trust alongside Neil Buddy Shah, Richard Fontaine, and Mariano-Florentino Cuéllar. Bernanke, who chaired the Federal Reserve from 2006 to 2014 and received the Nobel Prize in Economic Sciences in 2022, brought economic expertise the Trust specifically sought as Anthropic prepares its market debut.
The headcount dropped to three on August 4, 2026, when Cuéllar — a former California Supreme Court justice and former Carnegie Endowment president — stepped down from trustee seat to become Anthropic’s first Chief Global Affairs Officer. Three trustees: Shah, Fontaine, Bernanke are now listed on Anthropic’s official company page. The Trust is seeking a successor, but no appointment has been announced.
The timing is notable. The Trust is supposed to be at its most consequential as Anthropic finalizes the governance structure it will carry into public markets. A three-member body operating at reduced capacity — while simultaneously finalizing a new class of founder stock whose interaction with the Trust’s own Class T authority has not been publicly described — is asking public investors to trust a mechanism that is currently understaffed and in transition.
What Public Shareholders Will Actually Control
The combination of super-voting founder shares and Trust Class T shares produces a board that public investors cannot meaningfully change. That is not an accident; it is the explicit design.
Anthropic is organized as a Delaware Public Benefit Corporation, a legal form that allows — and legally protects — its directors when they trade shareholder returns for public benefit purposes. The LTBT’s Class T shares will, under the milestone schedule established in the company’s original charter, grant trustees the power to elect a majority of the board. The dual-class founder shares add a second, parallel restriction: even on matters where the Trust does not exercise its board-appointment authority, the founders’ enhanced votes would override the aggregate preference of ordinary public shareholders.
The Facebook case is the documented precedent for governance limits. When Zuckerberg held approximately 1% of Meta’s economic equity but retained roughly 60% of voting control through super-voting shares, shareholders who wanted to respond to the Cambridge Analytica data scandal found they had no path to removing the CEO or forcing material governance changes. The structure that protects founders from short-term shareholder pressure is the same structure that protects them from accountability for specific, documented failures.
Anthropic’s counter-argument is built into its PBC structure: it is legally organized to subordinate shareholder returns to mission. The Trust, in this framing, is the mechanism that makes that commitment credible and durable — not a mechanism that removes accountability, but one that redirects it to a different, non-financial standard. That framing is coherent. Whether institutional investors who have already pushed back against SpaceX’s dual-class structure — with three pension funds managing $1 trillion in assets publishing a letter calling SpaceX’s arrangement “the most management-favourable governance structure ever” — will accept the same tradeoff at Anthropic remains to be seen.
How Anthropic’s Safety Paradox Created the Control Problem
There is a specific irony embedded in the super-voting plan that the draft governance narrative tends to obscure. Anthropic’s founders did not lose their equity stake to ordinary market forces. They lost it specifically because the company’s mission required continuous, large-scale external capital to stay at the frontier — and at each fundraising round, safety commitments that made Claude credible to enterprise buyers also increased the investment terms that diluted the founding team.
The logic runs in a complete circle: safety-first positioning attracted enterprise revenue, enterprise revenue attracted massive investor capital, investor capital diluted founders, diluted founders now need super-voting shares to remain in control, and the stated purpose of remaining in control is to preserve the safety commitments that started the loop. The structure being built before the IPO is, in one reading, the governance equivalent of trying to hold a company’s original values in place against the very investors whose capital was required to get there.
That tension is precisely what Anthropic’s eventual S-1 — with its confidential S-1 under SEC review and not yet publicly available — will force into the open for the first time. Investors pricing a $2 trillion debut — the figure projected by six Anthropic investors speaking to FT — are pricing a company where they are explicitly buying into the upside while accepting that they have no lever on the downside.
What Does Sunset Provision Mean for This Structure?
Whether the founder super-votes will include a sunset provision — a time-based expiration date after which the enhanced votes revert to ordinary shares — has not been disclosed. For comparable tech IPOs, sunsets of seven to twenty years have been common in recent years. The Council of Institutional Investors specifically targets dual-class structures without sunsets as a governance risk, and ISS recommends votes against directors at newly public companies that have unequal voting rights without time-based expiration.
A sunset provision at Anthropic would interact unusually with the Trust’s own mechanism. The Class T shares that grant the Trust its board-appointment power are not subject to a sunset in any description published by Anthropic. If founder super-votes expired while the Trust’s Class T authority remained in place, the practical outcome would be a company where public shareholders remain unable to control the board — just for a different reason. That structural residue means that even the governance concession most commonly demanded by institutional investors at tech IPOs would not fully resolve the investor control question at Anthropic.
What Investors Have Not Seen Yet
The confidential S-1 under SEC review Anthropic filed with the SEC on June 1, 2026, has not been publicly released. When the public prospectus emerges — likely in late summer or early fall, on the current timeline targeting a late-September or October listing — it will contain the first audited financial statements, the full risk-factor disclosure, and the complete voting structure description that institutional investors need to make a grounded decision. Anthropic selected Goldman, Morgan Stanley, and JPMorgan as lead underwriters for the offering.
Until that document is public, the governance architecture being described in news reports is real — it has been confirmed by multiple independent outlets citing sources familiar with the matter — but its full operational details, including the vote ratio for founder shares, the sunset provision if any, and the specific interaction between founder-vote authority and Trust Class T authority, remain undisclosed. Investors who participate in the roadshow or buy at listing before reading the prospectus in full are making a governance judgment on incomplete information.
That is the condition Anthropic is working to resolve before shares trade. How the structure lands with institutional buyers who have already spent 2026 pushing back against dual-class governance at SpaceX — and with SEC reviewers who will require full transparency on voting mechanics before clearing the registration — will determine whether the governance architecture Anthropic has built to protect its mission becomes the IPO’s defining feature or its defining risk.
Frequently Asked QuestionsWhat is the Long-Term Benefit Trust, and who is currently on it?
The LTBT is a Delaware purpose trust that holds Anthropic’s Class T shares — a share class with no economic value but with the legal authority to elect an increasing number of Anthropic’s seven board members, ultimately reaching a majority. Trustees have no equity in Anthropic and are compensated only for their time. As of August 19, 2026, the Trust has three current members: Neil Buddy Shah, CEO of the Clinton Health Access Initiative, who chairs the Trust; Richard Fontaine, CEO of the Center for a New American Security; and Ben Bernanke, former Federal Reserve Chair and 2022 Nobel economics laureate. A fourth seat opened on August 4, 2026, when Mariano-Florentino Cuéllar left the Trust to become Anthropic’s first Chief Global Affairs Officer. No successor has been announced.
If I buy Anthropic IPO shares, what can I vote on?
The full details of the voting structure will appear in the public S-1 prospectus, which has not yet been released. Based on what has been reported: (a) founders will hold a separate class of stock with enhanced voting power that is designed to override ordinary shareholders on major decisions; (b) the LTBT’s Class T shares grant the Trust independent authority to elect the majority of the board, which is not subject to shareholder vote. That means ordinary public shareholders are likely to face a board they cannot change through voting and a management team with enhanced votes that ordinary shareholders cannot override. The specific vote ratio for founder shares has not been disclosed.
Does Anthropic’s dual-class structure include a sunset provision?
This has not been disclosed. Common practice at recent tech IPOs has included sunsets ranging from seven to twenty years, after which enhanced votes revert to ordinary shares. the Council of Institutional Investors specifically advocates for sunset provisions of seven years or less as a minimum governance safeguard for dual-class IPOs. Whether the Anthropic structure will include a sunset — and whether that sunset would interact with the LTBT’s own perpetual Class T authority — is not yet public information.
What happens to the LTBT’s board power when the company goes public?
The Trust’s Class T shares and board-election authority appear designed to survive the IPO — the structure is embedded in Anthropic’s charter, which requires consent from both the Trust and a supermajority of stockholders to amend. Public investors are expected to receive Class A common shares. The LTBT’s authority to appoint the board operates independently of the founder super-votes, meaning the two layers of restricted governance are parallel rather than sequential. For public investors, this means the combined effect is additive: neither the founder-vote layer nor the Trust layer can be removed by ordinary shareholder action alone.