Anthropic’s IPO Letter: 4 AI Risks and 22% of Equity Pledged

The letter Anthropic’s seven co-founders put at the front of their amended IPO prospectus, filed with the SEC on 17 November names four broad categories of AI risk, including that “Authoritarian governments may use AI to create high-tech surveillance states, deepen repression, and make themselves much more difficult to dislodge”.

Our number: the co-founders’ 80% pledges (9%), employees’ matched pledges (8%) and the Public Benefit Fund (5%) add up to 22% of fully diluted shares as of 30 June 2026, by our arithmetic from the letter and the filing.

Business Pill · Public Benefit Corporation

An evergreen Business Engineer explainer on how a public benefit corporation writes purpose into its charter. It teaches the general idea only and does not discuss this filing.

The key insight: Anthropic’s founders used the letter in their IPO prospectus to name four AI risks, from misuse to authoritarian surveillance, and to warn buyers they may slow down or invest heavily against investors’ expectations; pledges and the benefit fund equal 22% of fully diluted shares by our arithmetic.

The Four Risks the Founders Name

The letter says: “There are four broad categories of risk: misuse, loss of control, authoritarian abuse, and economic disruption”. On loss of control, it says “Losing control of sufficiently powerful AI systems could be a global security threat of the first order”.

On authoritarian abuse, the response is one sentence: “We are committed to working with free societies to provide a counterweight to authoritarian abuse and to promote freedom and democracy”. A page earlier, the same letter says Anthropic already works with “the U.S. federal government, many state governments, and U.S. allies in areas such as child services, social security administration, cybersecurity, and intelligence collection”. The authoritarian line drew attention on 18 November after ZeroHedge called it a “chilling warning”.

On economic disruption, the founders say they “have already committed $200 million to partner with governments to run studies on different ways of responding to changes in the economy”.

Anthropic's S-1/A puts co-founder pledges at 9%, employees' matched pledges at 8% and the Public Benefit Fund
Anthropic’s S-1/A puts co-founder pledges at 9%, employees’ matched pledges at 8% and the Public Benefit Fund at 5% of fully diluted shares as of 30 June 2026; together 22% by our arithmetic. Source: SEC filing.

The Warning to Investors

A section titled ‘On investor expectations’ quotes advice the founders say they received: “never let your company be worth a dollar more than it is actually worth”. It then sets out two ways Anthropic’s decisions “may run counter to investors’ expectations”: very large forward bets, especially in compute, that “may seem irresponsibly large at first glance”, and decisions that depart from profit maximization.

On pace, the letter says: “there may be times when we choose to advance more slowly than we could in order to make our systems safer, even if that puts us at a competitive disadvantage”. As an example it cites Claude Mythos Preview, which it says was held back from general release until new safeguards were built, a delay that “meant forgoing substantial revenue and ceding ground to competitors”.

What Buyers Get on Governance

The filing says the single Class F share held by the co-founders’ LLC carries 50.1% of the voting power on most stockholder matters and no economic rights. The letter says the Long-Term Benefit Trust elects four of seven directors, the co-founders elect the other three, and their extra voting power “passes to our common stockholders once fewer than three of us remain in the voting entity”. We compared founder voting power with other dual-class listings in our 12 November analysis.

The pledges behind our 22% are not all binding: the filing calls the co-founders’ 80% commitments “voluntary, revocable expressions of each co-founder’s philanthropic intent” that “do not constitute legally binding obligations”. We set the Public Benefit Fund against the founders’ stakes in an earlier piece.

The letter lands as Anthropic prices its offering: CNBC reported nearly $500 billion of orders and a possible upsizing, and safety researchers published a statement challenging Anthropic’s safety branding. Our running IPO tracker has the timeline.

The Structural Read

As we read the letter, it is written to pre-empt a future dispute with shareholders: it puts on record, before anyone buys, that safety-driven slowdowns and outsized compute bets are part of the plan.

As we read the filing, the 50.1% Class F vote is what makes that warning enforceable inside the company, whatever public stockholders prefer.

Anthropic’s seven co-founders, S-1/A founders’ letter

“there may be times when we choose to advance more slowly than we could in order to make our systems safer, even if that puts us at a competitive disadvantage”

Three Implications

FOR IPO BUYERS The filing gives the co-founders’ LLC 50.1% of the voting power on most matters through one Class F share.

FOR THE SAFETY DEBATE The letter names misuse, loss of control, authoritarian abuse and economic disruption as the four risk categories.

FOR GOVERNMENTS The letter says Anthropic already works with U.S. federal and state governments and U.S. allies, including in intelligence collection.

The Business Engineer Lens

This story maps onto the Business Engineer framework Inside Anthropic’s Permission Layer.

The analysis puts it this way: “It is becoming the first major AI lab actively attempting to define the governance layer of the AI stack.”

As we read the prospectus, the founders’ letter extends that governance stance to public shareholders. The analysis predates the filing and does not discuss it.

What Is Not Established

The letter does not say how Anthropic would decide when to slow down, who would make that call, or what it would cost; it states intentions, not commitments with numbers. Our 22% adds three separate figures the filing gives as of 30 June 2026 and mixes revocable pledges with shares already contributed; it is not a single figure Anthropic reports. We have not seen Anthropic comment on the ZeroHedge article.

This Week In The AI Supercycle

AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source

This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.

Read the full issue →

Business Engineer Framework

Inside Anthropic’s Permission Layer

A Business Engineer analysis of how Anthropic is positioning around the governance layer of the AI stack.

Read the Map of AI →

The Bottom Line

Anthropic’s founders used their IPO letter to name misuse, loss of control, authoritarian abuse and economic disruption as AI’s risks, and to tell buyers it may slow down or spend heavily against their expectations; pledges and the benefit fund total 22% of fully diluted shares by our arithmetic.

97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. The founders’ letter is printed as image pages at the front of Anthropic’s S-1/A Amendment No. 3 (accession 0001628280-26-074338); we read and transcribed those pages. Governance and pledge terms are from the filing’s text. The 22% is our arithmetic (9% + 8% + 5% of fully diluted shares as of 30 June 2026, as the filing states each). ZeroHedge is cited only as the source of the ‘chilling warning’ framing. We did not contact Anthropic.

Sources: Anthropic, PBC Form S-1/A Amendment No. 3 (SEC, 17 Nov 2026), founders’ letter and governance sections · ZeroHedge, 18 Nov 2026

Scroll to Top
FourWeekMBA