Anthropic Writes the Public Into Its Cap Table With a 5 Percent Pledge

According to Anthropic, 5 percent of its stock is to be given over to a Public Benefit Fund. The size of this pool is tied to the company's valuation and will expand as that figure goes up. In the case of a lab said to be in the process of staging one of the biggest public offerings on record, it means a portion of what is to come at the stock market is being put aside for the public good. What that money will be put to is the question on everyone's mind in Silicon Valley.

What Anthropic Announced, and What It Left Out

Anthropic set out the details of its new Anthropic Public Benefit Fund in an October 18 post called Investing in the public benefit. The fund is a fresh entity with a stake in the company that will grow in tandem with it. For now the commitment stands at 5 percent of Anthropic's shares, pending stockholder approval. In addition, the board has the option to put another 2 percent of shares into the fund for each 50 percent the company's value rises over a baseline figure which Anthropic has not made public.

The announcement did not go into much detail on spending. Anthropic said it will be putting out more information in the months ahead on the fund's governance and structure, as well as the valuation milestones, and a report on how it is faring with its mission will follow. The Verge observed that the company gave scant information on what would be done with the money aside from some illustrations.

They are of a general nature: Anthropic indicated the fund could make large donations to established nonprofits or direct distributions to specific populations, pointing to the labor market as one possible venue for action.

A Fund Built as an LLC, Not a Charity

The number is not the only thing of consequence, the structure is as well. The fund is a wholly owned limited liability company and not a 501(c)(3) foundation. Anthropic says this affords it the leeway to switch among various kinds of giving. A corporate philanthropy vehicle would not normally be set up that way and, in any case, it means the company holds sway over the equity instead of an independent board.

Then there is the hard line Anthropic has put in place. "None of the Anthropic equity held by the LLC will be used for lobbying or political giving, in the US or abroad," the company wrote in its announcement. In an industry given to heavy spending on policy battles in Washington and Sacramento, it is a boundary worth putting down in writing.

Not all public facing work will be transferred to the fund, some will be kept within the company. AI safety and interpretability research, applied biomedical research and the provision of discounted Claude access to scientific institutions and nonprofits are cited by Anthropic as having both commercial value and a public benefit. Those are activities that stay with the business.

Why 5 Percent Could Become 20 Percent

The new fund is in addition to what has already been put in place. From the day Anthropic was started, its co-founders made a commitment to give 80 percent of their equity, and the company has an equity donation matching program that roughly half of the staff make use of. With this fund included, Anthropic puts the figure at about 20 percent of its shares being set aside for the public good, not counting any further allotments from valuation milestones.

What is ingenious is the escalator. The fund is increased by 2 percent on every 50 percent rise in value, so the public's portion swells in step with the company's finest performance, when the money carries the greatest worth. Anthropic framed it as a pointed departure from the way tech firms typically reward success. "We are instead directing that equity to the public benefit," the company wrote, after noting that it is common practice to offer escalating compensation awards to executives.

The Timing Is Not an Accident

Anthropic is a public benefit corporation, and Reuters reports that it is planning one of the biggest IPOs ever. An IPO requires a company to put itself on the line before the broadest audience it can find; once a fund is put forward ahead of the final prospectus, it is part of that narrative. The 5 percent figure is for investors to see as an expense and for the public to take as a pledge.

The timing of the announcement is no accident given the growing skepticism over AI. Polls on data centers and suits brought over training data have been telling, not to mention the blockades on agents by banks and retailers this month. For a firm that expects its valuation to keep climbing, there is an interest in demonstrating that the climb is of some benefit to parties other than its shareholders.

The Labor Question Is the Real Test

Anthropic has put aside 200 million dollars already to study and address the impact of AI on the labor market and has identified it as the sort of problem the fund could handle with greater effect. The company wrote that "given the pace of AI progress, bolder approaches may be needed," and that the fund will have the versatility to pursue them.

As for what bolder entails, that is not defined. Direct distributions to specific populations could mean anything from retraining grants to cash for workers displaced by automation. The LLC set up for the fund makes such things possible but leaves the public to rely on Anthropic's judgment as to who is helped and in what measure, pending the governance particulars.

What Happens When AI Labs Start Paying a Public Dividend?

Should stockholders give the plan their approval, the first order of business could be a bout of competitive pressure. OpenAI, Google DeepMind, Meta and xAI might have to field questions from staff, regulators and customers as to whether they are prepared to put some of their own upside in service of public aims. A 5 percent pledge with an escalator offers a tangible benchmark for such comparisons and could spur a race to match it, putting actual money behind the labor and safety promises that have been little more than words of intent up to this point.

The model could also alter how AI firms are put together at the outset for founders and investors. A fund that expands with valuation is a matter for the cap table rather than the marketing budget; early stage boards may have to determine if equity is to be set aside for the public good prior to a Series A. Some investors might see value in taking a marginally smaller piece in return for a better social license, but others may nudge the founders into more traditional arrangements.

As for the workers, it all comes down to execution. Should the fund provide retraining or direct support at scale, it could rank among the first major private sector moves to counter job disruption from AI. Failing that, it may be seen as nothing more than a pre IPO gesture. The things to look out for in the months ahead are the promised governance announcement, the baseline on valuation and the inaugural report on public benefit, as well as any move by a peer lab to make a pledge of its own.

Putting Humanity First Is Also a Creative Question

Anthropic is putting forward the case that an AI firm should put some of its success in the hands of those the technology has an impact on. Sparknify has been looking at the matter from a creative angle with its Human vs. AI Film Festival and the Sparknify Humanity Award. This year the festival attracted more than 3,000 submissions from filmmakers the world over and on September 26 it premiered its 32 official selections in San Francisco.

The winning films had nothing to do with models or valuations. They were concerned with what one feels when a machine can put on a convincing act of memory, grief, humor or love and whether an audience is able to spot the distinction. It is the same tension Anthropic is seeking to address with equity: as AI does the things humans do with greater skill, the companies behind it must decide what they owe to the people affected.

A public benefit fund answers that in dollars; a film festival answers it in stories. Sparknify intends to follow both, since the coming years will be defined by the industry's treatment of people as much as by the pace of its model development.

Whether Anthropic's pledge becomes the new standard in the industry or stays a one off will tell you something about the type of AI company Silicon Valley is set on building.

Subscribe to Sparknify for Silicon Valley tech news and the people behind it.