Opinion

Anthropic donations: guesses & uncertainties

​tl;dr: before Anthropic IPOs, its staff won’t do large-scale philanthropy. If it IPOs, its staff will likely do CG-or-larger-scale longtermist philanthropy starting 3-9 months later. Anthropic is more likely than not to IPO late this year, but it might never happen.Anthropic recently sold equity at a $380B valuation. I think it would currently be valued at $600B if there was a real market for Anthropic equity, and you should expect it to grow ~80%/year in expectation (very volatile).Anthropic founders and staff are rich on paper, but they generally can’t sell their equity[1] until Anthropic IPOs.[2] Kalshi currently says Anthropic is 67% to IPO in 2026; this Manifold market agrees; in my view public information suggests being a little more bearish. If Anthropic IPOs at the end of the year, I expect it to be worth about $900B. Anthropic will likely impose a rule like “you can sell only a tiny amount of equity until six months after IPO” (after which you’re free to trade). On Anthropic staff’s DAFs, I don’t understand the situation but I think Anthropic or the DAF provider can set bad rules.In Anthropic’s recent tender offer, staff sold $5B of equity (no more than $10M per person, pretax). They also directed their DAFs to sell $250M of equity; their DAFs will soon have $250M. Staff would have liquidated more, but due to bad and changing rules (imposed by the DAF provider or Anthropic, I’m not sure), they were capped at $250M. (This is partially based on private info.)So around this week they’ll get like $5M cash each (post-tax, if they’ve been there for 2+ years), and $250M DAF liquidity collectively. That’s not much; almost all potential Anthropic philanthropy comes post-IPO.Here are my all-things-considered guesses about Anthropic ownership and donation propensities (disclaimer: quick guesses, like it’s impossible to overstate this, someone should actually investigate):Fraction of Anthropic owned by:Staff DAFs:[3] 20%Staff non-DAF (pretax): 6%Founders (pretax): 14%, of which 4/5 is pledged for donationCommunity investors (pretax): 5%: Tallinn 1-2%, Moskovitz 1%, Macroscopic 1%, other 1-2%Normie companies/investors: the remaining 55%[4]Propensities and cause prioritization of Anthropic founders and staff:Will it actually get donated? idk, some thoughtful people think that even post-IPO there might not be lots of good Anthropic philanthropy due to various messy dynamicsOf donated money: ~30% will go to decent longtermist stuff10% of $1T is $100B.If this is roughly correct, it seems a supermajority of longtermist community wealth is in Anthropic equity.[5]Note that you can get good investment returns. My AI-focused portfolio has returned some 65% per year (pretax) for the last 5 years. A certain hedge fund is substantially better. If you have money you should invest it well; this is important. If you want to put effort into this you should ask around for a google doc called “Thoughts on AI investments”; if not, you could do worse than just buying AIS. Not financial advice.In some ways it seems like there’s an opportunity for (formal but not contractual/enforceable) donation trades — if Alice has money to donate now, she could let Anthropic staff direct her donations, and post-IPO the staff would let Alice direct some of their donations. But trades like this between partially-aligned altruists are very cursed; you have to be careful.This post is part of my sequence inspired by my prioritization research and donation advising work.This post is very low-effort. It hasn’t been fact-checked. I feel good/stable about the first few paragraphs but the bulleted numbers may be totally wrong. If you have takes, feel free to DM me. But I’m shipping this in its current state because this kind of thing shouldn’t be a priority for me.^The longtermist community is overexposed to Anthropic equity; it would be better if the community could reallocate some Anthropic equity to other assets. Unfortunately selling equity is generally forbidden by employees’ and investors’ contracts, and the contracts also forbid creative approaches such as getting loans secured by your equity. Anthropic could permit sales but it isn’t inclined to.^Many people say “IPO” as shorthand for “become publicly traded,” even though you can go public without an IPO. I believe the “IPO” prediction markets would resolve Yes if Anthropic goes public without an IPO.^Actually on the order of 99% of this is just pledged for the DAFs, not yet owned by DAFs.^I think most people think this is more like 75%; I don’t know why but I suspect it’s just heuristics about companies which are incorrect or don’t apply in this case. Sorry for not justifying my beliefs; this post is low-effort and these numbers are wild guesses.^Note to self: if improving this post in the future, give context on:Current longtermist spending (i.e. mostly CG)Expected future spending flows (including OpenAI, SALP, and VARA)Discuss ​Read More

​tl;dr: before Anthropic IPOs, its staff won’t do large-scale philanthropy. If it IPOs, its staff will likely do CG-or-larger-scale longtermist philanthropy starting 3-9 months later. Anthropic is more likely than not to IPO late this year, but it might never happen.Anthropic recently sold equity at a $380B valuation. I think it would currently be valued at $600B if there was a real market for Anthropic equity, and you should expect it to grow ~80%/year in expectation (very volatile).Anthropic founders and staff are rich on paper, but they generally can’t sell their equity[1] until Anthropic IPOs.[2] Kalshi currently says Anthropic is 67% to IPO in 2026; this Manifold market agrees; in my view public information suggests being a little more bearish. If Anthropic IPOs at the end of the year, I expect it to be worth about $900B. Anthropic will likely impose a rule like “you can sell only a tiny amount of equity until six months after IPO” (after which you’re free to trade). On Anthropic staff’s DAFs, I don’t understand the situation but I think Anthropic or the DAF provider can set bad rules.In Anthropic’s recent tender offer, staff sold $5B of equity (no more than $10M per person, pretax). They also directed their DAFs to sell $250M of equity; their DAFs will soon have $250M. Staff would have liquidated more, but due to bad and changing rules (imposed by the DAF provider or Anthropic, I’m not sure), they were capped at $250M. (This is partially based on private info.)So around this week they’ll get like $5M cash each (post-tax, if they’ve been there for 2+ years), and $250M DAF liquidity collectively. That’s not much; almost all potential Anthropic philanthropy comes post-IPO.Here are my all-things-considered guesses about Anthropic ownership and donation propensities (disclaimer: quick guesses, like it’s impossible to overstate this, someone should actually investigate):Fraction of Anthropic owned by:Staff DAFs:[3] 20%Staff non-DAF (pretax): 6%Founders (pretax): 14%, of which 4/5 is pledged for donationCommunity investors (pretax): 5%: Tallinn 1-2%, Moskovitz 1%, Macroscopic 1%, other 1-2%Normie companies/investors: the remaining 55%[4]Propensities and cause prioritization of Anthropic founders and staff:Will it actually get donated? idk, some thoughtful people think that even post-IPO there might not be lots of good Anthropic philanthropy due to various messy dynamicsOf donated money: ~30% will go to decent longtermist stuff10% of $1T is $100B.If this is roughly correct, it seems a supermajority of longtermist community wealth is in Anthropic equity.[5]Note that you can get good investment returns. My AI-focused portfolio has returned some 65% per year (pretax) for the last 5 years. A certain hedge fund is substantially better. If you have money you should invest it well; this is important. If you want to put effort into this you should ask around for a google doc called “Thoughts on AI investments”; if not, you could do worse than just buying AIS. Not financial advice.In some ways it seems like there’s an opportunity for (formal but not contractual/enforceable) donation trades — if Alice has money to donate now, she could let Anthropic staff direct her donations, and post-IPO the staff would let Alice direct some of their donations. But trades like this between partially-aligned altruists are very cursed; you have to be careful.This post is part of my sequence inspired by my prioritization research and donation advising work.This post is very low-effort. It hasn’t been fact-checked. I feel good/stable about the first few paragraphs but the bulleted numbers may be totally wrong. If you have takes, feel free to DM me. But I’m shipping this in its current state because this kind of thing shouldn’t be a priority for me.^The longtermist community is overexposed to Anthropic equity; it would be better if the community could reallocate some Anthropic equity to other assets. Unfortunately selling equity is generally forbidden by employees’ and investors’ contracts, and the contracts also forbid creative approaches such as getting loans secured by your equity. Anthropic could permit sales but it isn’t inclined to.^Many people say “IPO” as shorthand for “become publicly traded,” even though you can go public without an IPO. I believe the “IPO” prediction markets would resolve Yes if Anthropic goes public without an IPO.^Actually on the order of 99% of this is just pledged for the DAFs, not yet owned by DAFs.^I think most people think this is more like 75%; I don’t know why but I suspect it’s just heuristics about companies which are incorrect or don’t apply in this case. Sorry for not justifying my beliefs; this post is low-effort and these numbers are wild guesses.^Note to self: if improving this post in the future, give context on:Current longtermist spending (i.e. mostly CG)Expected future spending flows (including OpenAI, SALP, and VARA)Discuss ​Read More

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *