Everything About AI
NextFin News - Anthropic is preparing for an initial public offering that investors have discussed at as much as $2 trillion, and the deal will ask public shareholders to do something they have never been asked to do at this scale: accept a board majority elected by outsiders who own no shares.
The San Francisco artificial-intelligence company confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission on June 1, 2026, giving itself the option to go public once the regulator completes its review. The number of shares and the offer price have not been set. What is already visible is the governance architecture Anthropic intends to take public: a Public Benefit Corporation with three share classes, founder super-votes, and a Long-Term Benefit Trust whose members hold no equity yet carry the authority to elect a majority of the company's seven-member board.
The tension is stark. At a valuation that would make it the largest U.S. listing ever, Anthropic is not merely asking investors to pay for growth. It is asking them to accept that the people with the most direct power over the board are three trustees who receive no dividends, capture no capital appreciation, and cannot be removed by a shareholder vote.
The Structure: An Inverted Model of Corporate Control
Most technology companies that go public with restricted voting ask investors to trust the founders. Alphabet did it in 2004, when its founders controlled 51% of the voting rights while owning less than 12% of the shares. Meta did it in 2012, when Mark Zuckerberg retained 61% of the voting power with 13% of the economic ownership. In both cases, control and money sat in the same hands: the insiders who owned the company also ran it.
Anthropic has separated the two. Its capital structure has three tiers:
- Class A shares, which public investors are expected to receive, carry one vote per share and standard economic rights.
- Class B shares, held by co-founders Dario and Daniela Amodei and select co-founders, carry super-voting power. The exact vote ratio has not been disclosed and will appear in the public prospectus. The super-vote plan was first reported in August 2026 and represents the first time Anthropic's leadership would hold enhanced voting power.
- Class T shares, held exclusively by the Long-Term Benefit Trust, carry no economic value at all. Their only purpose is governance: the authority to elect and remove designated directors.
The trust's power is not static. It was designed with an escalating schedule: elect one director initially, then two, and eventually three — a majority of a five-member board — within four years or upon hitting certain fundraising milestones. As Anthropic has scaled, that authority has expanded with it. The trust now holds the power to elect a majority of a seven-member board, and that authority is embedded in the company's charter. Amending it requires consent from both the trust and a supermajority of stockholders, meaning ordinary shareholders cannot unwind the structure even if they unite against it.
The current board reflects the trust's reach. It includes Dario Amodei, Daniela Amodei, Yasmin Razavi, Reed Hastings, Chris Liddell, and Vas Narasimhan. Reed Hastings was appointed to the board by the LTBT in May 2024, and Jay Kreps was appointed by the trust in April 2024 as its first elected director. The founders run the company. The trust protects the mission. Public shareholders sit beneath both layers.
Why Anthropic Built It: The Safety Brand as an Asset
The company's own description of the trust is explicit about its purpose. "The Trust's arrangements are designed to insulate the Trustees from financial interest in Anthropic and to grant them sufficient independence to balance the interests of the public alongside the interests of Anthropic's stockholders," the company has said.
That framing is not incidental to the IPO — it is the product. Anthropic has spent years marketing itself as the safety-conscious alternative in frontier AI, the lab that would not cut corners on alignment to win a product race. The trust is the institutional embodiment of that claim. By placing real board power in the hands of people who cannot profit from the stock, Anthropic is arguing that its governance is what makes its safety promises credible.
The roster is built to reinforce that argument. The trust was announced in September 2023 with five founding trustees drawn from AI safety, global health, and effective-altruism organizations. Since then it has churned: Jason Matheny stepped down in December 2023 over potential conflicts with RAND Corporation's policy work; Paul Christiano left in April 2024 to become head of AI safety at the U.S. AI Safety Institute; Kanika Bahl and Zach Robinson concluded their terms in January 2026. Richard Fontaine, CEO of the Center for a New American Security, joined in May 2025. Mariano-Florentino "Tino" Cuéllar, president of the Carnegie Endowment for International Peace, joined in January 2026. Ben Bernanke, the former Federal Reserve chair and 2022 Nobel laureate in economics, joined in July 2026.
Then came the complication. Cuéllar left the trust on August 4, 2026, to become Anthropic's first Chief Global Affairs Officer — the first publicly disclosed instance of a trustee moving into the company itself. The trust, which stood at four members as recently as July 9, 2026, is now down to three: Neil Buddy Shah, Richard Fontaine, and Ben Bernanke. For a structure whose entire selling point is independence, a trustee crossing over to management is the kind of event investors will scrutinize.
The Numbers Behind the Ask
The trust question does not sit in a vacuum. It arrives alongside financials that explain why this IPO matters and why the valuation is so aggressive.
Anthropic's last private mark was $965 billion, set on May 28, 2026, when it closed a $65 billion Series H round. Investors have discussed a public valuation of up to $2 trillion — roughly double that private mark — which would make it the largest initial public offering in history, surpassing Saudi Aramco's $29.4 billion raise in 2019. Reports have discussed a raise of $60 billion or more, with Morgan Stanley, Goldman Sachs, and JPMorgan leading the underwriting, and a listing that could come as soon as October 2026.
The revenue trajectory is what makes those numbers discussible rather than absurd. Annualized revenue moved from about $9 billion at the end of 2025 to $14 billion in February 2026, $30 billion in April, $47 billion in mid-May, and $65 billion at the end of July — a path that doubled the run rate roughly every six weeks at points, according to people who have reviewed the company's figures. In the second quarter of 2026, the company reportedly projected its first positive operating result, with operating profit of about $559 million on revenue of about $10.9 billion, excluding stock-based compensation. Cash-flow breakeven, by contrast, is not expected until 2028.
The $2 trillion figure does not rest on today's revenue. It rests on a projected 2028 revenue range of $190 billion to $200 billion. At $2 trillion against $200 billion of 2028 revenue, the company would trade at 10 times forward sales. At the $965 billion Series H mark against the $65 billion July run rate, it already traded at roughly 15 times trailing revenue. The public-market case, in other words, requires investors to believe the growth rate holds even as the revenue base becomes large enough that percentage gains should naturally slow.
Market leadership is the other pillar. In April 2026, Anthropic overtook OpenAI in U.S. business AI spending for the first time — 34.4% versus 32.3%, according to spending data — and leads enterprise large-language-model API usage at roughly 40%, ahead of OpenAI's 27% and Google's 21%. The company says about 80% of its revenue comes from more than 300,000 business customers, and that the number of customers spending $1 million or more annually has topped 1,000, up from more than 500 in February.
Its biggest backers are locked in but locked out. Amazon and Alphabet hold stakes reported at roughly 21% and 15% respectively, and both are contractually barred from board seats or voting rights. Amazon, for its part, has already booked enormous paper gains: in the second quarter of 2026 its net income surged 245% year over year to $62.6 billion, driven largely by its Anthropic investment, which lifted its "other" net income from $1.1 billion to $53.4 billion — about two-thirds of the quarter's pre-tax income.
"Today, Anthropic, PBC confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission for a proposed initial public offering of our common stock. This gives us the option to go public after the SEC completes its review," the company said in its announcement.
The Counter-Thesis: Is External Trust a Feature or a Discount?
The strongest argument against Anthropic's structure is that it asks investors to pay a premium for a governance mechanism that may not work as advertised, while stripping them of the normal tools to hold management accountable.
Critics point to the trust's actual record. One published critique of the trust put it this way: "The LTBT has seemingly done very little thus far—appointing only 1 board member despite being able to appoint 3/5 of the board members (a majority) and the LTBT is down to only 3 members. And none of its members have technical expertise related to AI." The critique cuts two ways. If the trust has been passive through the company's fastest growth, investors must decide whether that reflects prudent restraint or an ineffective check. And as the roster shrinks and a trustee moves into management, the independence claim takes on water.
There is also the question of price. Public markets have historically discounted companies with restricted voting. The Council of Institutional Investors has long advocated for sunset provisions of seven years or less on dual-class structures, and investor groups have repeatedly challenged arrangements that separate voting from economic ownership. Anthropic's structure is more extreme than a standard dual-class: it is a dual-class plus an external veto layer. If investors demand a governance discount, the $2 trillion aspiration becomes harder to reach even if the revenue story holds.
And there is a deeper, structural objection: the trust's power is perpetual while the founders' super-votes may one day sunset. The company has not disclosed whether the Class B shares carry a sunset provision. If founder votes eventually revert to one-share-one-vote while the trust's Class T authority remains, public shareholders could end up with a board majority they never elected and cannot remove — a permanent external bloc with no equity at risk.
The counter-argument, and Anthropic's likely response, is that this is precisely the point. A trust that could be outvoted or unwound by shareholders would not credibly constrain the company in a moment when safety spending conflicts with quarterly results. The structure is designed to be hard to change. The question is whether the market treats that durability as a safeguard or as an entrenchment device.
This is the crux of the cyclical-versus-structural call. A cyclical governance problem — a weak board in a strong quarter — would self-correct once shareholders vote or the stock price forces a change. Anthropic's design removes that correction mechanism by intent. If the structure works, it is a structural shift in how public companies can be owned: control detached from capital, permanently. If it fails, the failure will not be a cycle that reverts; it will be a precedent that discourages the model for a generation. There is no middle ground where the market simply waits it out.
What Would Prove the Thesis Wrong
The bull case for Anthropic's governance rests on a testable proposition: that an independent, non-shareholder board bloc produces better long-term outcomes than conventional governance, and that investors will pay for that difference rather than discount it.
Two signals would falsify that proposition. First, if the trust remains at three members — or continues to lose members to the company — through the IPO and the first year as a public company, the independence narrative weakens materially. Second, if the shares price at a persistent discount to comparable high-growth AI names after the initial trading window, the market will have answered the governance question with its wallet: external trust, in this structure, is a risk to be priced, not a premium to be paid for.
Regulatory scrutiny is the third variable. The SEC's 2026 examination priorities explicitly target AI-driven threats to data integrity, third-party vendor risk, and enhanced board oversight of AI governance as part of cybersecurity risk management. A December 2025 letter from 42 state attorneys general warned AI companies about "sycophantic and delusional" outputs and urged stronger safeguards. Anthropic's disclosures around how the trust actually exercises its board authority — not just that it exists — will matter as much as the structure itself.
Outlook: Three Scenarios for the Listing
Base case. Anthropic lists in late 2026 at a valuation between its $965 billion private mark and the $2 trillion aspiration, with the trust structure intact and a board majority traceable to the trust. Trading is volatile but the revenue growth story carries the stock, and the governance question fades into the background as long as quarterly results beat. Short-term sentiment is driven by AI-sector positioning; medium-term fundamentals hinge on whether the $190 billion to $200 billion 2028 revenue target stays credible.
Upside case. The IPO prices at or above $2 trillion, the trust is framed successfully as the reason Anthropic can be trusted with frontier models, and the stock rerates higher as enterprise adoption continues to outpace OpenAI's. In this scenario, Anthropic's governance becomes a template that other AI labs copy, and the "external trust" model becomes a category of its own.
Downside case. Investors demand a governance discount, the trust's reduced roster and the Cuéllar transition become focal points in the roadshow, and the stock opens below the $965 billion private mark. That would be a bruising signal for the entire private AI valuation complex, where trillion-dollar marks have been set without any public-market test.
Across all three scenarios, the time horizons point in different directions. In the short term, the IPO will be a liquidity event and a sentiment barometer for the AI trade. Over the medium term, the question is whether Anthropic can keep doubling its revenue run rate as the base grows. Over the long term, the real experiment is whether a company can be publicly owned while its board is controlled by people who have no ownership at all.
Anthropic's IPO will not just test whether the market will pay $2 trillion for AI growth. It will test whether public investors are willing to hand the keys to a board they cannot fire, and whether a trust with no money at risk can be trusted more than shareholders with everything at risk. The shares are the secondary offering. The trust is the primary one.
Explore more exclusive insights at nextfin.ai.
