Anthropic's Benefit Trust Has Picked 4 of 7 Directors Ahead of a $2 Trillion IPO, FT Reports
A.I. / news
Anthropic's Benefit Trust Has Picked 4 of 7 Directors Ahead of a $2 Trillion IPO, FT Reports
The 2023 design gave the trust one seat of five and a majority within four years; the FT's September numbers put it at four of seven.

Anthropic's Long-Term Benefit Trust, which holds almost no equity, has already chosen four of the company's seven directors, according to a Financial Times report from September 4, as the company prepares an IPO that could value it at up to $2 trillion. The design Anthropic published in 2023 promised the trust a board majority within four years. The FT numbers suggest it is nearly there.
What the trust is and what it holds
Anthropic described the trust in a September 19, 2023 post. It holds a class of stock, Class T, that lets it elect and remove some directors. "In any event, the Trust will elect a majority of the board within 4 years," the post says.
Trustees "serve one-year terms and future Trustees will be elected by a vote of the Trustees." The trust had five members at the outset.
A Harvard Law School forum post dated October 28, 2023, by John Morley of Yale Law School with David Berger and Amy Simmerman of Wilson Sonsini Goodrich & Rosati, who served as Anthropic's outside counsel, adds the detail on money. The Class T shares carry a financial claim the authors call "very small."
Then and now
The two descriptions are three years apart and come from different kinds of source, so the table sets out which is which.
| 2023 design (Anthropic, Harvard forum) | 2026 report (FT via Dealroom) | |
|---|---|---|
| Directors the trust elects | 1 of 5, rising to 3 | 4 of 7 selected |
| Trustees | 5 | 3 in place, 5 possible |
| Equity | "Very small" claim | None |
| Named trustees | Not listed | Neil Buddy Shah, Ben Bernanke, Richard Fontaine |
The 2026 column comes from Dealroom's summary of the FT's Big Read, not from the FT itself, which was not available to read. Dealroom attributes the interpretation to the FT's Murgia.
The equity row is a discrepancy. The 2023 forum post says the claim is very small; the 2026 summary says the trust holds none. Neither source defines the difference.

What the trustees see before launches
The FT report, as summarised by Dealroom, says trustees get advance notice of major actions, including new model launches. They meet weekly among themselves and with leadership as often as every other week.
Dealroom's summary says trustees encouraged a limited rollout of Anthropic's Mythos cybersecurity model through the Glasswing Project. That is the one concrete example of the trust affecting a product decision in the sources read.
How it could be removed
The FT said, per Dealroom, that the trust can be removed with the support of 85% of shareholders' voting power. The Anthropic post says only that stockholders can change the trust "without the consent of the Trustees" if large enough supermajorities agree, with thresholds that rise as the trust's power phases in.
Dealroom's summary does not say whether the 85% figure applies at the IPO or whether public shareholders would vote under the same rule. Founders' voting rights are not covered in the summary either.
What is still unknown
The IPO timetable and valuation are not settled. The FT figure is that Anthropic "may be valued at as much as $2tn." No source read here gives a filing date.
For other capital-markets and model coverage on this site, see Tenstorrent's $1.5 billion Series E and Gemini 4 Argon's launch.
Anthropic's 2023 commitment puts the latest date for a trust majority at September 2027. The IPO prospectus, once filed, is where the current numbers would have to appear.
Sources
More in A.I.
- 01RuneBench Has AI Agents Grind RuneScape Skills, and Its Leaderboard Puts GPT-6 Astra on Top at 7.3%The benchmark's own page publishes no scores, and the aggregator that does says the numbers mix model quality with harness choices.
- 02OpenAI's Rosalind Academic Plan Gives US Labs 30 Free Seats for 12 Months, Without API AccessApplications opened Oct. 15, and the model itself still sits behind a separate trusted-access review.
- 03Disney Offers Jobs to 60 Character.AI Staff in Bankruptcy, and Reports Differ on the CountA court filing prices the hires at about $60 million a year, which works out to roughly $1 million a head.
- 04Claude's Scorpion Toxin Binders: 17 of 211 Designs Bound, and None Has Been Tested Against VenomAnthropic's Oct. 20 report counts one toxin with no hits and says the work stops short of an antivenom.