Dear Reader,
Yesterday was our most well-attended webinar in our company’s history.
I just want to thank you for that.
We are deeply humbled that you take time out of your schedule to spend any of it with us.
We know you have a million options out there, a million people you could talk to or listen to.
We view this as a tremendous responsibility to you, our customers.
We are eternally grateful.
I just wanted to start by saying thank you.
Today I want to talk about Anthropic.

You saw the earnings, but even putting that aside, I thought about it yesterday.
I think this whole Huang debacle is delivering a great stress test for the market, and for the AI story in particular.
That happens in bull markets.
There’s always something that gives you a reason to sell.
Investors have to look past it.
They have to come up with an argument that gets them past it.
And as you get earnings like last night, the story advances, people get more confidence, and they go out and buy.
As I’ve said a million times, I don’t view the AI story as over at all.
It’s short-term turbulence.
Turbulence on the plane, not a lost engine.
That’s just not the way I see it.
No way.
Absolutely not.
Anthropic Prices Today
Anthropic is going public.
It prices today, which, if you’ve worked on Wall Street, you know happens late in the day.
It starts trading Thursday.
Pricing a security is a very complicated process.
You’re hearing from all these investors and taking an order book.
Twenty thousand shares for this guy, fifteen thousand shares for that guy.
It’s very old school.
They have a lot of tools to make it new school.
But at the end of the day, there’s a person trying to get the most for the company and the best price for investors.
The symbol is ANT.
The Bull Case
I made a bit of a bearish argument the other day.
My concern is durable competitive advantage, and I don’t know if this company has one.
But there are always two sides to every case, especially with a company like this.
So today I’m going to give you the bull side.
They’re talking about bringing it public in a $1.25 to $1.4 trillion range.
Several people on the Street say it will be worth a little more.
Some believe that valuation is cheap, and here’s why.
OpenAI just did a raise at $1.25 trillion.
Anthropic’s business is bigger than OpenAI’s.
It did $17 billion in the third quarter, versus OpenAI’s $10 billion.
That’s a business 70% bigger at a valuation only about 10% higher.
That doesn’t seem to make sense.
Now consider this.
Say Anthropic ends the year at roughly a $100 billion annualized revenue run rate.
That’s a bit above the $81 billion people have.
Say it doubles that over the following two years.
Now apply a 25% adjusted operating margin.
It did 19% in the June quarter.
As revenue grows, per-unit costs fall and margins expand, so 25% is reasonable.
On that math, Anthropic could earn roughly $50 billion in operating profit in 2028.
Wow, $50 billion.
That means the proposed valuation of $1.2 to $1.5 trillion implies a multiple of over 25 times 2028 EBIT, which is earnings before interest and taxes.
At a mid-twenties EBIT multiple, Anthropic would be kind of cheap relative to software multiples.
A little richer than some mega cap tech, and it’s pushed out a bit.
The Real Question
But the real question isn’t whether Anthropic works in the short term.
It can go up 20%.
It can go down 10%.
Who knows?
That depends on how the order book is handled, and on whether people are feeling euphoric that day.
Mr. Market is in play here.
People have been really negative on this stuff the past year, ever since the midterms.
The AI trade’s over, and all this other nonsense.
So it depends on how it opens.
The valuation maybe goes higher and maybe goes lower.
Maybe it’s worth $1.2 trillion, and maybe it’s worth $1.7 trillion.
Reasonable people can make arguments for all of those cases, truly.
I’ll tell you one thing, though.
I’m glad it’s coming down from that $2 to $2.25 trillion valuation, because that just didn’t seem reasonable.
They brought it down 20%, 30%, 40%, which is definitely more reasonable.
If you’re going to play around, that’s the kind of place you want to play.
In the short term, the valuation question isn’t driven by ultimate free cash flow.
It’s driven by whether this can become another Google.
Can it get to be the size of another Google or another Microsoft?
That’s the question you have to ask yourself.
And you know what?
Probably.
The Pricing Power Problem
The question is whether it will have pricing power.
In a world where the product you make can be copied so cheaply by China, what real advantage do you have?
Say I patented a drug that saves your life, and I’m selling it for $50 a pill.
Then all of a sudden somebody can copy it and sell it for $1 a pill in the same market.
I’m not sure I understand what happens.
I don’t know how your prices stay up.
I don’t know how they don’t go down over time.
Of course, the US government could say, national security, we’re not going to let Chinese models into the United States.
Fair enough.
These models have a lot of advantages stateside.
But internationally, they’ll be running on their own stack.
So I ask myself how you defend your pricing power.
Everybody is coming out with revenue projections.
They’re basing them on current pricing power, not on a really open, competitive market.
Bottom Line
We’ll see.
There’s more to be revealed.
Lord knows I’m not the guy with all the answers.
I gave you the bear argument the other day.
I wanted to give you some bull argument today, for those of you who’ve been asking whether I’d recommend buying Anthropic.
I’ve been asked a hundred times.
To me, whether it goes to $1.7 trillion or $1.25 trillion, who cares?
Maybe it goes up 20%, maybe it goes down 20%.
That’s not what this game is about for Anthropic.
You either believe this thing is going to be another Google, or a Meta, or an Apple, or a Microsoft, or you don’t.
If you actually believe that’s going to happen, then you buy it.
Have a wonderful day.
I’ll see you tomorrow.
“The Buck Stops Here,”

P.S. Everybody wants to know whether Anthropic will have pricing power.
There’s another way to play AI where that question doesn’t really come up.
A little-known chip-design company collects a royalty on the chips inside Tesla, Starlink and SpaceX’s new AI satellites.
It doesn’t have to win a price war.
It gets paid on every unit built.
And SpaceX has filed with the FCC to launch up to one million satellites to run AI data centers in orbit.
That could multiply those royalties roughly 100 times.
The window I’m watching closes December 8th, when the SpaceX insider lockup expires.
Found this helpful? Share it with others.
Written by Dylan Jovine