Circus.
That’s one word (probably the best one) to describe Anthropic’s IPO roadshow, which launched earlier this week ahead of what, if everything goes according to plan, will be a Nasdaq debut for the ages just before Thanksgiving.
On Wednesday, reports indicated the offering, which includes a private placement, is “multiple times” oversubscribed despite a lot of Sturm und Drang across the AI space in recent weeks.
All in, Anthropic could raise nearly $90.5 billion, enough to make it the largest IPO in history.
There’s the chart. It’s worth noting that when you include the underwriters’ overallotment option, SpaceX raised $85.7 billion
As part of the anticipatory carnival, Anthropic’s live-streaming a fire-side-style chat with Dario Amodei (who, as far as we know, didn’t sleep with Xi Jinping) and Jamie Dimon (who hopefully didn’t either). The two of them will discuss the significance of the first public equity offering from a frontier lab and AI more generally.
Must be nice. Right? I mean, I’m no investment banker, but as far as I know, the average company hoping to go public can’t just hop on Dimon’s Calendly and pick a date for a televised one-on-one marketing event.
The spectacle, set for noon on Friday, will be the second time in six months that Dimon’s lent his celebrity to a blockbuster IPO. This week’s Anthropic event will be “similar to but less extravagant,” as Bloomberg put it, than the spectacle Dimon put on for Elon Musk earlier this year.
In June, Dimon personally pitched the SpaceX deal at JPMorgan’s headquarters. That event, The New York Times recalled, was broadcast to dozens of the bank’s branches and offices around the country, some of which “closed early” for the occasion. (Bangs on branch door. “What do you want?” asks the security officer. “I need to deposit this check and get some cash back,” says the customer. “Tough shit, we’re closed.” “What? Why?” “We’re watching the SpaceX event. The ATM’s full service, though. You can do everything you need to do there.”)
At the high-end of the price range announced by Anthropic, the company would be worth more than $1.4 trillion.
As the figure shows, that’d make it larger than uncle Warren’s insurance premium-financed hedge fund — sorry, sorry — “conglomerate.” It’d also be more valuable than Walmart. And Dimon’s bank. Not too shabby for a company that’s all of — checks filing — 70 months old.
As mentioned here last week, the AI jitters which tripped up tech stocks since mid-October have adversely impacted Anthropic’s prospective valuation. Some potential investors reportedly cut their estimates to as low as $1.1 trillion this month, down from $1.5 trillion just weeks previous. Those estimates were as high as $2 trillion late in the summer.
The figure below gives you some context for the company’s sales run rate, which extrapolates full-year revenue from quarterly sales.
Anthropic’s amended S-1 (linked above) tips $81 billion on that metric.
“This shows that the stabilization created by the release of Opus 5.5, in response to OpenAI’s Astra, is holding up, but it does not show acceleration,” JonesTrading’s Mike O’Rourke remarked.
Meanwhile, Goldman sees a lot of upside. The firm’s equity analysts are modeling enterprise-client revenue of $586 billion by 2030 and another $100 billion from “regular” users. That’d be $685 billion in all, up almost 150x from last year’s sales.
This really is unbelievable and I mean that in a good way. I have no idea whether or not $586B is realistic, but the fact that I have to think about it and can’t dismiss it out of hand says a lot.
I expect semis will come back down to earth once their capacity catches up to demand, but I’m not sure Anthropic will.
My daughter builds systems for a Silicon Valley startup serving large pharma customers. She’s told me she uses Claude daily in her work. I don’t know what that costs her company, but it seems to make her more efficient. Her latest new customer involves 26 product lines at Pfizer.
I can’t go back to writing software without AI. Like, if it goes down, I’m taking a walk or doing something else (not work) until its back.
Absolutely incredible. I pay $200/month. I would personally pay $2k/month. And my company would pay a lot lot more. I do the work of 10-20 people without AI.
I can’t stress it enough.
Yeah, I’ve just eclipsed Bloomberg / WSJ / Reuters / FT’s entire graphics / data visualization departments in one relatively inexpensive OpenAI upgrade. Of course, my following isn’t big enough for them to care. I’m not talking about me, specifically. I’m just saying this is an existential moment for Bloomies and LSEGs, even if they don’t know it yet. Those things are antiquated. Already. By this time next year, they’ll be total dinosaurs. The only reason you’d need a terminal right now is if you had to be 1,000% sure that your calculations were “right” and that the data was derived as much from first-hand / primary sources as possible. If you have any leeway or margin for error there at all — and by that I mean if your business doesn’t depend on every, single tick being “justifiable” by the whole “Well, I pulled it from a Bloomie!” fall back — then you don’t need a Bloomie or LSEG anymore. I’m gonna keep my access, but only for the news feed. I don’t need the rest of it.
And there is a good reason why MSFT took a hit yesterday too. Yeah, Azure is a big part of their valuation, but a Claude office suite is a threat in a much bigger way than the google office suite ever was. I’m still giddy about Claude data. That feels like my version of what Claude Code and Codex were for programmers.
Truly unbelievable that revenue figure and whilst there are some strong returns to be had for specific use cases, $586 billion revenue equates to $1,700 per person pa!