November 21, 2026
Heard in AI

Moonshots panel splits over Anthropic's trust ahead of its IPO

On Moonshots, Salim Ismail called Anthropic's Long-Term Benefit Trust a radical way to build its mission into governance. Alexander Wissner-Gross warned it could slide into ideological control as public shareholders buy in.

A briefing reports one development at a point in time. We may correct or clarify it later; a new development gets a new briefing. How our formats work

Based on Moonshots with Peter Diamandis, episode published October 28, 2026 (recorded October 27, 2026)

Salim Ismail, founder of Open ExO, read Anthropic's public S-1 for a question other than the revenue: who gets to choose the company's board. An S-1 is the registration statement a company files with US regulators before it sells shares to the public, and Anthropic filed its own on October 26, 2026. On the Moonshots podcast, recorded the next day, Ismail called Anthropic's Long-Term Benefit Trust "a pretty radical form of corporate governance." Alexander Wissner-Gross, a computer scientist and founder of Reified, saw in it something that reminded him of the way a ruling party controls companies.

The panel did not settle the question. The conversation became a debate about who should steer a company that Peter Diamandis, the show's host and founder of XPRIZE, said "will be one of the most powerful companies on the planet for quite some time."

What the trust does

Anthropic is a public benefit corporation, a legal form that lets a company pursue a stated public purpose alongside profit. Since September 2023 it has also had the Long-Term Benefit Trust. According to Anthropic's explanation of the trust, the trust holds a special class of stock, Class T, that lets financially disinterested trustees elect and remove directors. That power grows as time passes and funding milestones are reached, until it covers a majority of the board. The company calls the arrangement an experiment meant to balance public benefit against shareholders' interests. Anthropic's board picked the first trustees, and the trustees elect their own successors. An October 2026 update lists the current trustees as Neil Buddy Shah, Richard Fontaine, Ben Bernanke and Christopher Stubbs.

The trust is one of two control mechanisms in the S-1. The filing gives a founder-owned LLC a single Class F share carrying 50.1% of general voting power, while Class T separately governs the election of the board majority that the trust appoints. In practice the founders would hold most of the shareholder vote, and the trust would choose most of the directors.

Ismail: the mission gets "the power to choose the table"

Ismail said it was "something I've never seen before." He described it with MTP, short for massive transformative purpose, his term for the ambitious mission an organization sets for itself. Anthropic, he said, had "embedded the MTP into the governance."

Diamandis brought up Larry Page and Sergey Brin, Google's co-founders, who kept control of their company after it went public. Ismail said Anthropic's setup "goes even further," because control does not rest with the founders alone. It turns the mission "from a motivational statement into a governance mechanism," he said. "The mission, the MTP gets a seat at the table and the power to choose the table."

He also named the test he expects the trust to face: "what happens when doing the right thing costs you $100 billion? Who makes that trade-off?" And he raised the old question of "who watches the guardians": who decides what is good for humanity, and who holds the trustees accountable. "But directionally," he said, "I find this totally fascinating."

Wissner-Gross: a route to "ideological control"

Wissner-Gross said the financials were "great," but the Class T stock left him "quite concerned." He contrasted Anthropic with OpenAI. In his telling, OpenAI started as a nonprofit and "sleepwalked" into public benefit corporation status through lawsuits and deals with Microsoft, while Anthropic was a public benefit corporation from the start. His objection was about Anthropic becoming a listed company. He said it would be ceding control of potentially most of its board seats "voluntarily, preemptively" to a trust of independent people who are "maybe not necessarily economically aligned," taking influence over major decisions away from ordinary Class A shareholders.

He offered an analogy that he called "slightly overstrained": the way the Chinese Communist Party exerts ideological control over Chinese companies, with a committee "ostensibly motivated by certain ideological considerations" overriding the economic interests of the wider public. Founder control is different, he said. Founders created the company, are economically tied to how it does, and "it's their baby." The trustees, as he understood it, are people the founders bring in because they "perhaps" share the founders' views. His fear was that the arrangement would "degenerate into ideological control."

Later he sharpened the image. Every generation may be "doomed to repeat the mistakes of the past and reinvent all the isms" of the early 20th century, he said. He worried that an institution like this could become "a smoke-filled back room where decisions on the future light cone are made." The light cone is a physics term for everything a present event can affect in the future. He set that room against "the entire market" helping to make those decisions.

Wissner-Gross framed the handover of power as something that happens when Anthropic lists. In fact, Class T and the phased board powers have existed since 2023, according to Anthropic's own explanation. That design also includes amendment procedures and failsafes that let a supermajority of stockholders make changes, with thresholds that rise as the trust's powers grow. What the listing changes is that public buyers would be accepting an arrangement already in place.

Blundin: shareholders want grit

Dave Blundin, founder and general partner of Link Ventures, judged the trust the way an investor judges a management team. He pointed to the hardship Steve Jobs and Jensen Huang went through: Jobs "getting fired, humiliated, coming back," and Huang "on the brink of bankruptcy multiple times." "That grit is what a shareholder looks for in a management team," he said, "not the ideological, you know, fluffy stuff." Anthropic's path from founding to success had been "so short and so pain-free," he said, that this would be an interesting test for its leaders. He did not think the Chinese Communist Party comparison was "out of bounds."

Blundin did not doubt the motives of Dario Amodei, Anthropic's chief executive. Anyone who thinks Amodei is in it for the money "has lost their mind," he said; Amodei is "100% in it for the good of the world." He asked whether that should scare a shareholder and answered that "it might." He called it "the one weakness" and said he agreed completely with Wissner-Gross.

Ismail: circular appointments and the market's "high metabolism"

Ismail called Wissner-Gross's point "very valid." Because the trustees appear to appoint one another, he said, the structure is circular, which could lead to groupthink at the trustee level and "a total lack of accountability." Anthropic's own description bears out the circular part: after the first trustees, the trustees elect their successors.

He still liked the direction. The trust brings in "general civilization stakeholders," he said, and not only people who own capital. "We just have to make sure they're the right people," he added, and not "communist leaning Marxist types." He also explained why Anthropic might not want markets in charge. Markets run on "very short-term high metabolism" earnings cycles, he said, and companies driven by them "often don't do the right thing long-term." Anthropic was trying to balance that pressure, and whether it struck the right balance, he said, is still to be seen.

A vote for the models, and the case for staying private

The discussion then moved to two more radical ideas. One was that Anthropic should give each of its models a vote on the trust, on the reasoning that the models are now capable enough to play that kind of role.

The other was that Anthropic should never have become a public company at all, if its stated mission is taken seriously. By the figures offered on the show, Anthropic has raised $130 billion, holds about $110 billion in cash and cash equivalents, and has spent only $10 billion to $20 billion of it. Another $100 billion raise would leave around $200 billion, comparable to the cash piles of Apple and Google.

Diamandis asked why Anthropic was going public, suggesting it needed the cash. Blundin answered with one word, "Compute," the processing power needed to train and run AI models. "There's no amount of compute that's enough," he said. The reply was that Anthropic does not need the cash: it would not care about its share price, would tap the markets whenever it needed to, and would find that easier as a public company, but at the cost of conflicts it could have avoided by staying private. Blundin answered, "No doubt."

The argument ended on a contrast between two styles. Elon Musk, it went, will work to boost his company's share price, while Amodei will say the share price does not matter because the work is for the good of the people, and Anthropic might even stop holding quarterly earnings calls. That, in this view, is a misalignment on a fiduciary basis, meaning the legal duty a company's directors owe its shareholders, and a reason Anthropic should have stayed private through the next few years of recursive self-improvement, when AI systems increasingly help build better AI.

Share this article

Go to the original

Sources & further reading

  1. 01
  2. 02

Connected ideas and articles

From the conversation

Podcast episodes

Moonshots with Peter Diamandis

Anthropic’s $2T IPO, Elon Predicts Neocloud Death, and China’s Kimi Catches Up | #306

Episode published (recorded )This article draws on 37:04–45:14 (approximate times)

Article history

Updates to this article

Tags