# Anthropic IPO: Valuation, Governance, and AI Risk

**Podcast:** Unhedged
**Published:** 2026-10-27

## Transcript

Will it be the biggest IPO ever?
Will it be the biggest thing ever?
Might it destroy the world?
Today on the show, Anthropics IPO.
This is Unhedged, the markets and finance podcast from the Financial Times.
I am Robert Armstrong coming to you from a heavily improvised recording studio deep underground at Unhedged World headquarters in New York City.
I am joined by the famously clever John Foley.
Welcome to the show, John.
Thanks for having me, Rob.
We've got an enormous document yesterday, the prospectus, the S1.
And it's almost 400 pages long or something.
It weighs as much as a cement block, something like that.
What did we learn from this enormous document, John?
Well, Rob, we learned that Anthropic, which is the AI lab that makes the clawed chatbots, is planning to go public.
It's going to be a listed company probably at some point before the end of November.
We learned that it's growing really fast.
Its revenue is growing like nearly 1,400% rates year-on-year, which is like a lot, going like 15 times bigger year-on-year.
We learned that it's going to spend a lot of money on things like building data centers and doing research into AI labs, which are likely to sap its profit and certainly its free cash flow for many years.
And we learned that it thinks it might possibly incinerate everything that we love through some kind of apocalyptic AI mishap.
Good times.
You wrote a great LexNote about this yesterday, which basically made the argument that the appropriate thing to do with this massive and highly detailed Bible of an S1 is to throw it away and ignore it.
Could you walk us through that rather contrarian argument?
Yeah, print it out, throw it in the trash can.
The reason for this, I'm being flippant, and I did not do this, I did actually read it.
The reason for this mostly is that...
As an IPO investor, you're trying to work out what the company's worth.
And I think that the rearview mirror numbers that you get in this filing are of almost no use in doing that because they show a company that is changing so quickly, didn't even exist before 2021 and hadn't made any revenue for 2023.
And to try and extrapolate where those numbers are going to go in the very long term is really a fool's errand.
I agree with you.
I just want to scale how big some of these numbers that don't really matter or tell us very much about the future are.
First half revenue, first half of this year, 16.3 billion American dollars, like 15 times more than it made in the first half last year.
So one thing we do know is that there's a lot of money sloshing around here.
In the second quarter of the year, that revenue was growing 1,368%.
So the point is, this company is growing really quickly.
And that makes its historic numbers, which are what we get in this prospectus, useless.
for calculating what the company is going to be worth.
Are you suggesting it's not going to grow at 1,000% indefinitely, John?
Well, if it grows, I just, at back of napkin, was like, what if it does grow at this rate for the next five years?
And I worked out that it would make something like $584 trillion of revenue by 2031, which is not possible, I don't think.
I've always thought the Lex column, a good other name for it would be the back of the envelope column.
They're like fantasy maths, Colin.
Basically what we do is scribble things on the back of bits of paper.
That's what we do.
Cigarette packets.
And the one thing that is really useful, which is what management thinks it's going to make in revenue in future years, they don't put in the filing.
They will talk about it on their roadshow.
How on earth would they know any better than us?
Well, that is a great question because what Anthropic is going to make in revenue and profit, which is what you care about if you're a shareholder, depends on how quickly people adopt their AI products, how good those products are.
are, what their rivals do, whether there's a price war, how much they pay for their data centers.
The variables are innumerable and every one of them change because the company is growing so quickly.
It changes the valuation dramatically.
If you tweak a knob just a tiny bit in one direction, the valuation changes considerably.
So take your pick.
I think what's really going to happen is they will just pick a number, which is the number they have in mind, which is probably about a trillion and a half.
And everyone will scramble to justify it.
Yeah, do the math backwards to get to a trillion and a half.
So the numbers in this massive document don't help us all that much beyond the point of saying, boy, that number is big and it is growing fast.
Exactly.
Right.
But let's talk a little bit about the rhetoric in this document, which I thought was remarkable.
I'll start with a piece I...
particularly liked.
This is a sentence from the early part of the prospectus.
The AI exponential will generate so much value that in the long run, it will far outweigh any short-term economic sacrifices we make as part of our mission.
And that kind of captured for me the spirit of the whole thing, that the wave that we're riding is so great that however much we spend on the surfboard, it's just not going to matter in the end.
And a couple of things struck me about that.
One, it's reminiscent of IPO documents I read in the past from companies like Meta and Google, where that basically turned out to be true.
They gave very similar warnings.
We're going to spend a lot of money.
It's all going to be worth it.
And by God, it was.
So that's point number one.
Point number two I'd say about it is there's something deeply ambiguous about this statement.
And what I mean by that is that this exponential, and by the way, I love using exponential as a noun, the exponential.
That's very science fiction.
I love it, actually.
Will generate so much value.
Well, value for who?
Like I think we're all convinced that AI as a technology will generate a lot of value.
The question is, for who?
Right.
Is it going to be captured by consumers?
Is it going to be captured by this company, another company, a whole bunch of companies?
So many companies that no individual company ends up making a lot of money.
It seems to me that's the whole question of this prospectus.
Yeah.
I think like what you when you describe that, the exponential, I mean, it does have a bit of a messiah complex ring to it, doesn't it?
This idea that.
the world is going to change and it's going to be because of us.
It may look like we're just going to incinerate hundreds of billions of dollars of your money, but...
We are going to change everything.
And actually now you get to be part of the Maasai complex too because you can put your money into this amazing project which will make everything amazing for everyone.
And they say this.
They say at the very end of this very high-flying introduction, they say, we want you to think of this as being part of something bigger.
Right, it's philanthropy.
And this just makes me want to puke.
It really does.
I hate that.
And, you know, we can come back to this theme.
But I'm like, just be a company, guys.
Right.
You know what I mean?
We're a democracy.
We will deal with all this stuff.
You're filing for an IPO.
Everybody just calm down a little bit.
You know what I mean?
Yeah.
My view is that one of the absurd things about this and the other reason you should throw the filing into the dustbin is that it is, we're all trying to pretend that this is a normal company and that you can force Anthropic through the sausage machine that we use to make publicly traded shares in companies.
But this company is not normal.
We don't know how to value it.
It's spending crazy amounts of money on things that we don't know how that's going to come out.
It has, as I'm sure we're going to talk about, a very unconventional governance structure.
And it might destroy everything.
How can you possibly price a company that is going to destroy everything?
Up until your last sentence, I completely disagree with you.
This is a company.
It is going to have a CEO.
The CEO is going to have lieutenants.
It's going to have salespeople.
It's going to sell products.
It's going to have a finance department.
It is a very fast-growing company.
There's plenty of fast-growing companies out there.
There's huge companies out there.
All this is totally normal.
The hard question is your final point, which is I would summarize it like this.
Would you have put the Manhattan Project into an IPO?
Would it have been appropriate?
to take Oppenheimer and the gang and Niels Bohr and whoever else was working over there and say, you know what the right structure for this project is?
We're going to IPO it.
And so you're raising the question, maybe that analogy does apply here.
And if their rhetoric about the existential possibilities of this technology is true.
Maybe it shouldn't be a publicly traded company at all.
I guess there are two questions though.
One is like, should anyone be allowed to put money in this thing that might destroy everything, which is your Manhattan?
project question, really, right?
Like, should we just starve this and all companies like it of money to protect the species?
Or, but there's a separate question.
Or just give it to the government.
Or the UN, or, you know, the Sisters of Mercy.
Feels unlikely at this point to happen.
But the other question is just, is more, like, in the libertarian spirit of the US capital markets, sure, list every, just disclose everything, do what you like.
If investors will pay for it, they'll pay for it.
But, like, let's not pretend that you need to fill 350 pages with facts when you could write this in a tweet.
Like the pitch for Anthropic is we're going to spend loads of money.
We might change the world and everything in it.
Do you like it or do you not like it?
As I already intimated, I kind of take the other side.
I think like a very standard corporate structure can contain this and that the mechanisms that are going to keep us safe from the worst permutations of this technology will always exist outside of the company.
And we have to work on those things.
And plenty of companies produce very, very dangerous things.
And the world knows how to handle those situations, you know, noxious chemicals to nuclear power, you know.
So we disagree.
But one thing I want to discuss here is the governance structure of this company, which is unique.
First of all, it is a public benefit corporation.
Can you give our listeners a sense of what that means?
This is the non-lawyer version of how this works.
So please send in your complaints if you're a lawyer listening to this.
To katie.martonft.com.
So normally if you're a shareholder in a company, the company has to do what's best for you, the shareholder.
And if they don't, you can sue them and claim that they breached their fiduciary duty.
With a public benefit corporation, the company is set up in such a way that it doesn't just have to do what you, the shareholder, want.
It's allowed to say, we are actually serving three groups.
We're serving you, the shareholders.
We're serving general stakeholder communities that come into contact with us.
And we're also following a purpose that we decide.
And in this case, Anthropik's purpose is to, and I'm paraphrasing, but it's basically to responsibly make and maintain advanced AI for the benefit of humanity.
So it means if you are a shareholder and you're angry with what Anthropik's done, because they've incinerated hundreds of billions of dollars in data centers that turned out to be worthless, you can't sue and say, This company didn't do what was best for me because the company was, well, we didn't have to just look at your interests.
We are a public benefit corporation, so the rules are a bit different.
And courts have found quite recently that companies do have leeway to do what's not best for shareholders if they're considering these other groups too.
So that's unusual.
There are some companies that have this structure that are listed, but not many and none as big as this.
So under that broad aegis of a public benefit corporation, there's a very specific structure here.
which I think of as kind of two centers of power within this company.
One center of power is the seven founders.
And like in many big Silicon Valley startups that got really, really big, it's set up so the founders have effective voting control of the company to the degree it's a normal company, to the degree it's a company with a board and shareholders can vote on actions and what the board has to do or whatever.
These seven founders have 50.1% of the voting shares.
This is something that could happen in any kind of company.
And so they have a locus of control.
The other locus of control is in something called the Long-Term Benefit Trust, which is a group of five very wise people.
I suppose, or purportedly.
Can you walk us through a couple of these names?
So it's going to be five.
It's currently four because there's an empty chair, which is wild to me that they've started this process before they've filled that seat.
Are you waiting for the call, John?
They haven't even, and also, by the way, there are seven board members and they've only got six at the moment.
They still haven't found seven.
Two jobs.
By the end of the month, we could both be employed.
One each.
I'll do the board.
What they've done is they've picked four people and there will be a fifth.
And they are sort of like dignitaries, experts in their fields whose job is to kind of say like – well, who are in theory responsible for overseeing anthropics' duty to protect humanity.
The one for like finance and economics folks, the one who's going to stand out is Ben Bernanke who's the Fed chair.
Let's just pause on that for a second.
I love the idea that we're taking potentially the most dangerous technology.
in the history of our species and handing it to the man who invented quantitative easing.
He didn't necessarily invent it, but he was the guy who made it big.
He's seen some stuff.
Yeah, he's...
No, no, I'm not saying he's a bad choice, but it's, you know, this is a finance show, so we had to put a little asterisk to that.
And they have, like, they've...
There are, you know, four people on that board.
They represent...
you know, finance economics through Bernanke.
And then there is a public health person.
There is a national security person.
And there is a like a science physics.
I'm being very like broad brush with these terms, but like science physics.
It's like a superhero team.
They got Superman.
They got Batman.
They got Aquaman, you know, Wonder Woman's on there.
But these people, while experts, their job, they are not managing the company day to day.
They don't have any say over the company day to day.
All they do is nominate people to sit on the board who then...
oversee the management of the company.
And this is where the rubber hits the road.
It's a seven-member board that runs the company, and the LTBT trustees get to nominate four of them.
The shareholders, a group of people which, as we've noted, is controlled by the seven founders, they get to nominate three.
So in theory, if the two sides get in fight, it's four to three, and the long-term...
benefit trust, I keep having to look up the order of those words, wins.
It's a tug of war.
But I think on balance, I think the trust, the four-person trust, has the upper hand in that.
I would point out that one of the trust-nominated directors at the moment is Reed Hastings, who was the co-founder of Netflix, who was famous at Netflix for having horrible governance practices.
He was a great leader of Netflix.
Under him, Netflix ignored shareholder votes.
It kept directors who'd been voted out after they'd been voted out by shareholders.
He was not an adherent to conventional corporate governance, and he is now one of the trusts.
Here's my takeaway, John, from the governance.
They've set it up almost like they're setting up a country, that they're the authors of a constitution.
It feels like it's like Montesquieu or John Locke is sketching out something where there's this delicate balance of power.
with all these rather obscure and picky tie-breaking mechanisms, which ultimately end, by the way, if you follow the rabbit hole all the way down, it comes to a vote of the employees.
So it literally kind of goes to the electorate in some weird way at the end.
And again, this reflects the great or possibly grandiose vision of what this company is that the founders have.
This has to be governed like a republic that is going to lead the world.
And, you know, there's a lot of democracy talk in the prospectus.
So it's another point about the tension we've been talking about.
You know, whereas I read this and think, God, guys, just treat it like a company.
Your view seems to be more like this is a special case and some of this stuff makes sense.
So I would say two things on the governance, really.
Again, we don't know if it really will destroy the planet.
Anthropoc is saying that it might.
And I think, as Dr.
Maya Angelou said, when someone shows you who they are, you should believe them the first time.
That's my view.
But like the two, one thing to know is that as a regular shareholder, you have no say over anything.
Like all of this stuff that we're discussing happens like far removed from you.
You don't get a vote, apart from in very rare circumstances.
The other thing, though, is that all this mechanism that we're describing between the trust and the founders and the board nominations.
will happen quite slowly.
If there's a problem, it will take quite a long time for that to translate into direct pressure on the founders.
But in AI, things go wrong very quickly.
Look at the stuff we've been seeing out of OpenAI, hacking, hugging face, the Australian healthcare system.
Things will go wrong in a lightning flash.
And this governance structure is great over the decades, but I don't think it's going to save us if something goes really wrong.
By way of closing.
I would like to talk about what we think this company is worth as distinct from what the market is going to decide it's worth, which I agree with you is going to be whatever it says it's worth.
If they and their bankers say they want $1.5 trillion, they're going to get their $1.5 trillion, and that's how it's going to go.
For me, the problem is, as I intimated before, I have no idea what the competitive structure of this market is going to be.
And therefore, I have no idea what the margin.
this company is going to be.
And so there's a big blank space.
So you have to value it as a kind of lottery ticket kind of story, right?
I think that's partly right.
I think when we say things like the numbers are useless, valuing it's pointless, I mean, it sounds like we're just abdicating our responsibilities as financial journalists to do some work on it.
And that isn't the case.
We've been through...
Applying multiples of revenue, for example, which you can easily get to 1.5 trillion just by valuing it like SpaceX.
You get to a trillion if you put it on a revenue multiple along with Meta platforms, which means Facebook.
I think that part of it is, and this is uncomfortably unscientific, but people will look at it and say, what is the stature of this company?
Is it like a SpaceX?
Is it like a Meta?
Is it like an alphabet?
Is it that important?
And if it is, its valuation will be roughly what theirs are.
SpaceX is $2 trillion, right?
Anthropic is, we're looking at like $1.5 to $2 trillion maybe because it feels like a SpaceX.
Like what is Meta?
Meta is about $1.8 trillion.
That is the current going rate for a big company that we feel is important.
And I hate that because I want to see numbers and analysis.
We will be right back, listeners, with Long and Short.
Listeners.
Welcome back.
This is long and short, which is, as you know, that portion of the show where we go long things we like and short things we don't like.
John, what do you like or not like today?
Well, I wouldn't say I like it, but I'm going to go long plane tickets.
Regretfully long plane tickets, because I think the cost of air travel, which has already gone up significantly here in the US, is going up further.
Oil prices are wreaking havoc on airlines, and this week we're hearing that JetBlue...
which is a budget airline, is in trouble and its creditors are organizing in preparation of what might be some kind of restructuring and the bonds are trading at like 40 cents on the dollar.
So no more JetBlue would mean stronger position for the big airlines, which will put prices up even further.
So I think even when the oil price comes down, we're looking at some long-term changes to the price of Gage and M&A.
Well, thanks for that piece of good news, John.
I'm really sorry to tell you this.
Book your tickets now, people.
I'm going to go short hats today.
And in general, I like hats.
As you know, I write about clothes sometimes and I approve of hats, but I think they're very dangerous items.
And what brings this to mind is a story that is in the FT right now about Donald Trump Jr.'s trip to Mongolia for hunting.
And the only thing I can remember about this story, having read through it, is the picture of the president's son wearing an enormous, absurd fur hat.
And for me, this picture kind of like...
captures everything that is ridiculous about the current administration.
And I just think if you are in political life or a public figure at all, be very careful about what you put on your head.
It's very cold in Mongolia.
I have been there.
You have.
And I can tell you it gets very cold, yeah.
But you've got to think, you know, it's like, you would call this the Mike Dukakis rule.
You may not remember this.
I do remember this.
He was, like, doing well in his run for president, and somebody put an army helmet on him, and it was all over.
Yeah.
And this hat, I'm telling you, it could do it for Donald Trump Jr.'s political aspirations.
Listeners, we will be back in your feed on Thursday wearing very sensible headwear.