Imagine owning a slice of one of the most valuable companies on the planet and handing it to a fund meant for everyone else. That’s roughly what Anthropic said it would do on Sunday, 18 October. The company behind the Claude AI assistant plans to put 5% of its shares into a new Anthropic public benefit fund. And by one estimate in CNBC’s coverage, that slice could be worth at least $50 billion.
So is this the biggest social finance move of the year, or a clever bit of PR before a stock market debut? Honestly, it might be both. Let’s break it down.
So what did Anthropic actually announce?
Here’s the short version. Anthropic wants to move 5% of its equity into the Anthropic Public Benefit Fund, a new entity it’ll fully own. The plan still needs a thumbs-up from stockholders, though, as AP reported.
The fund isn’t a classic charity. Instead, Anthropic is setting it up as a wholly owned limited liability company, or LLC. In other words, it’s a company structure, not a nonprofit. That choice matters, because it gives the board room to move money in different ways over time.
Anthropic also framed the move as a choice about where value goes. Rather than piling ever-bigger pay packages on executives, the company said, “We are instead directing that equity to the public benefit.”
One more detail stands out. The fund can’t spend money on lobbying or political donations, in the US or abroad, according to Al Jazeera. That’s a smart guardrail, since an AI company funding politics would raise eyebrows fast.
How the Anthropic public benefit fund can grow
This is the clever part. The 5% is only the starting point. For every 50% jump in Anthropic’s value, the board can add another 2 percentage points to the fund. So the better the company does, the bigger the public’s slice gets.
In Anthropic’s own words, “the more Anthropic grows, the larger the share of the company that can be dedicated expressly to the public good.”
Then there’s the bigger picture. Add this fund to earlier pledges, and roughly 20% of Anthropic’s shares will go to the “public good,” AP says. CEO Dario Amodei and five other cofounders have already pledged to give away 80% of their wealth. On top of that, there’s an equity donation match for staff, and about half of employees take part.
For context, Anthropic is a public benefit corporation registered in Delaware. That status means its leaders have to weigh investor returns against responsible AI work. So this fund fits the brand. Still, a pledge like this is easy to announce and much harder to run well.
Who’s the money actually for?
Anthropic hasn’t picked a single cause yet. But it’s dropped some big hints. In its blog post, the company said “we might explore large donations to traditional nonprofits or direct distributions to specific populations.”
That last phrase is the interesting one. Direct distributions means sending money straight to groups of people, instead of only to charities. And one cause the company has named is the disruption AI could bring to jobs.
Anthropic has already committed $200 million to studying AI’s effects, including what happens to the labour market. Yet it says the pace of AI progress may call for bolder approaches. This fund gives it room for those bolder moves.
The worry is real, too. A McKinsey report from last month, cited by Al Jazeera, put a number on it. Up to 11 million American workers may need to change careers over the next decade because of AI automation. That’s a lot of people looking for a new path, and a lot of pressure on whoever claims to help them.
Okay, but what’s the catch?
Here’s where it gets less rosy. The fund stays inside Anthropic, and the company’s board decides where the money goes and how much it gives away. CNBC’s Kate Rooney noted that “some are questioning why Anthropic would retain control of the money versus just giving it away directly.”
There’s also a math problem for future shareholders. Because the fund grows as the valuation climbs, new investors give up part of their upside. As 24/7 Wall St put it, “the better the stock performs, the more new shareholders lose to the fund.” Nobody knows yet whether the extra equity will come from new shares or from existing holders.
Then there’s trust. Outsiders “cannot check how or when a fund the company controls actually pays out,” the same analysis argues. Rooney also said the pledge “does appear to be a move to win some public trust around AI.” Fair enough. AI isn’t winning many popularity contests right now, with job fears and fights over data centres.
Anthropic’s answer is that control lets it pace donations sensibly, while the bans on lobbying and political spending limit self-dealing. But the full governance and payout terms still haven’t been spelled out.
Why should impact investors care?
At EIF, we spend a lot of time thinking about how capital can do social good. This story is a neat case study, because it blurs the line between a company, a foundation and an investment fund.
Most impact investing starts with choosing where money goes. This flips it. Anthropic ties a share of its future gains to public benefit, so the fund’s size depends on how well the business performs. If you’ve read our explainer on social impact investing, you’ll spot the tension we always come back to: how do you prove the outcome, rather than the intention?
That’s also why the transparency questions matter so much. Regulators already want funds to back up their impact claims, as we covered in our piece on ESMA’s impact guidance. A corporate fund this big will face the same question sooner or later.
For emerging markets like Armenia, the open question is reach. Will “specific populations” mean only American workers? Or could some of this money flow to places where AI’s effects will land differently? We’d love to see it travel.
So, should you cheer or squint at the Anthropic public benefit fund?
Both, honestly. A pledge that could be worth tens of billions is good news, and the growth rule is genuinely creative. But a fund that the giver also controls needs clear payout rules before anyone hands out medals.
Here’s what we’ll watch. First, how the board decides who gets money, and whether outsiders get a clear view of it. Next, whether Anthropic adopts binding payout rules, as its critics want. Then, whether the fund ever reaches beyond the US.
For anyone working in social finance, the lesson is pretty simple. Big pledges are a start. But the people a fund claims to help deserve to know what it’ll do, when it’ll do it, and how they’ll know it worked.
So here’s a question for you, and we’re genuinely curious about the answers. If a tech company promised you a share of its future success, what would you want in writing first?