Anthropic Put Public Benefit on the Cap Table

Anthropic just made one of the more interesting philanthropy promises in AI.

Not because the number is large.

Because the number is a percentage.

On October 18, the company announced that, subject to stockholder approval, it will commit 5% of Anthropic's shares to a new Public Benefit Fund.

The board may add another 2% every time Anthropic's value rises 50% above a baseline that has not yet been disclosed.

The fund could make large donations to nonprofits. It could also make direct distributions to specific populations. Anthropic points to workers affected by AI as one possible focus.

This is not a normal corporate donation.

A fixed check gets smaller relative to the company that wrote it.

An equity stake grows with the company.

If AI creates the extraordinary value its builders keep forecasting, some of that upside would already belong to a vehicle created to distribute public benefit.

That is the right instinct.

It is not yet a complete system.

Five percent is not twenty percent

Anthropic says roughly 20% of its shares will be dedicated to the public good when the new commitment is combined with earlier pledges.

That sentence needs careful accounting.

The new fund does not own 20% of Anthropic.

It is due to receive 5%, subject to stockholder approval.

The rest comes from different commitments: Anthropic's co-founders pledged to donate 80% of their own equity holdings, and about half of employees participate in an equity donation and company-matching program.

Those promises may all support public purposes.

They are not the same asset, held by the same entity, under the same rules.

The additional 2% steps are not automatic either. Anthropic says they will be added at the board's discretion when valuation milestones are reached.

Until the company publishes the baseline, the milestones, and the transfer mechanics, the clean number is five.

Conditional five.

Equity is not money

A share can become extremely valuable.

It cannot pay a worker's rent until somebody can turn it into cash.

Anthropic is establishing the fund as a wholly owned LLC rather than a charity. The company says this will let it move between different forms of giving, including direct distributions, and that the LLC's Anthropic equity will not fund lobbying or political donations.

Flexibility can be useful.

It can also hide every operational question inside the word “flexibility.”

When may the fund sell shares?

Who decides how much to sell and when?

Will the stake be protected from dilution in future financing rounds?

Can Anthropic repurchase it?

What happens if the company's value falls sharply before a distribution?

Does a private valuation trigger a new grant even when the fund still has no liquid market for the shares it already holds?

An equity promise is a claim on future value.

A public-benefit program is a way to deliver value to people.

Anthropic has announced the first. It still has to design the second.

Three structures, three different jobs

Anthropic already has two public-purpose structures.

Its status as a public benefit corporation gives directors legal latitude to balance shareholder returns, the interests of people affected by the company, and Anthropic's stated mission.

Its Long-Term Benefit Trust holds a special class of stock and has the power to elect and remove an increasing share of Anthropic's board, ultimately a majority. That is a mechanism for corporate control.

The new Public Benefit Fund is supposed to hold economic value and distribute it.

These are three layers:

• The PBC defines what directors are allowed and required to consider.
• The Trust influences who controls the company.
• The Fund is meant to move some of the company's gains into public hands.

One cannot substitute for another.

Legal permission is not accountability.

Board control is not distribution.

And a distribution fund does not give its intended beneficiaries a voice over the company creating the disruption.

That final gap matters. A recent Columbia Law School analysis of Anthropic's governance describes it as the separation of benefit and control: the people a public-benefit corporation says it serves generally receive no corresponding governance rights.

Anthropic's fund could narrow that gap.

But only if workers, communities, and independent experts help determine what “benefit” means in practice.

Philanthropy is not policy

There is a tempting story here.

AI companies create enormous wealth, reserve a percentage for everyone else, and privately manage the transition they accelerated.

That is not a social contract.

It is a company program.

Even a generous fund cannot replace taxes, labor law, unemployment insurance, retraining rights, antitrust enforcement, or democratically decided income support.

The company should not be both the source of disruption and the final authority on what affected people deserve.

This does not make the fund meaningless.

It makes its boundary important.

Private philanthropy can experiment faster than government. It can fund neglected research, test distribution mechanisms, support communities before legislation catches up, and produce evidence for better policy.

It should not become the argument for avoiding that policy.

The next disclosure needs receipts

Anthropic says it will publish the fund's structure, governance, and valuation milestones in the coming months.

That disclosure should answer at least seven questions.

Is the initial 5% transfer final and legally difficult to reverse?

How will the stake be treated during dilution, a sale, or a public listing?

Who governs the LLC, and how many decision-makers are independent of Anthropic?

Do intended beneficiaries have seats, votes, consultation rights, or an appeal process?

What rules govern liquidity and annual distributions?

How will conflicts between Anthropic's interests and recipients' interests be handled?

Will the fund publish audited holdings, grants, direct payments, administrative costs, and rejected proposals every year?

The answers will determine whether this is a durable institution or a large promise with a flexible wrapper.

Putting public benefit on the cap table is still a meaningful move.

AI companies usually ask society to trust that future abundance will spread on its own.

Anthropic has attached part of that future to an actual asset.

Now it has to make the promise legible, irreversible, and spendable.