画像クレジット: Image via source article. Used under fair use for news commentary. · source
Most of the coverage you'll read today about Anthropic's IPO filing focuses on the horse race angle. Who filed first! Who's worth more! Will Claude or ChatGPT win the AI wars!
Call me old-fashioned, but I think everyone's missing the actual story here.
Anthropic filed confidentially with the SEC on Monday, and yes, technically that puts them ahead of OpenAI in the race to go public. The Verge and Bloomberg both ran with the rivalry framing, which, fine, it makes for good headlines. But if you've been covering tech long enough (and I have, unfortunately for my blood pressure), you recognize a different pattern here. This isn't about beating OpenAI to some finish line. This is about the money clock running out.
Anthropic's post-money valuation after last week's fundraise hit $965 billion. That's billion with a B. OpenAI, for comparison, sits at $852 billion. These are the two most valuable startups on the planet, and neither one of them is profitable, and both of them are burning cash at rates that would make a 1990s dot-com blush.
I've seen this movie before! The late-stage private fundraising, the confidential S-1 filing, the "as soon as this fall" timeline that Bloomberg keeps repeating. This is what happens when the venture money starts getting nervous and the growth-at-all-costs model needs a new source of fuel. Public markets are that fuel.
関連記事
More in AI Models
I spent a week parsing the claims around Google's new 'always-on' AI agent, and the answer is more complicated than the marketing suggests.
Aisha Patel · 6 hours ago · 7 min
The AI company is now officially the world's most valuable startup, and it's moving fast toward public markets.
James Chen · 7 hours ago · 3 min
Everyone's writing about the $200B CPU market grab. The actual story is how Nvidia is quietly becoming the landlord of global AI compute.
James Chen · 7 hours ago · 4 min
The new superchip targets laptops and desktops, though what this means for edge robotics compute remains an open question.
The filing was confidential, which means we don't know the share count, we don't know the price range, we don't know the actual financials. What we do know is that Anthropic chose to file now, in June, with a potential debut "as soon as this fall." That's an aggressive timeline for a company that, let's be honest, most retail investors couldn't explain to you beyond "they make Claude."
Here's what strikes me as important:
Anthropic is a Public Benefit Corporation, which adds complexity to the IPO structure that nobody's really talking about yet
The confidential filing means they can test the waters with institutional investors before revealing anything publicly
"As soon as this fall" could easily become "sometime in 2027" if market conditions shift (and they always shift)
Neither company has disclosed what their actual path to profitability looks like, because, well, there probably isn't one yet
The young founders in AI seem to think trillion-dollar valuations are normal, that this is just how things work now. Maybe they're right! But I remember when WeWork had a $47 billion valuation right up until the moment it didn't, and I remember when Pets.com was going to revolutionize pet food delivery. I'm not saying Anthropic is either of those things, but the pattern of "file before the music stops" is familiar.
A confidential S-1 is basically a dress rehearsal. Anthropic gets to share their financials with the SEC and get feedback before anything goes public. They can revise, they can adjust their narrative, they can figure out how to explain to Wall Street why a company that's maybe two years old (depending on how you count) is worth nearly a trillion dollars.
Bloomberg's coverage mentioned that demand for Claude is surging, which, sure, I believe that. Claude is genuinely good, I use it myself (don't tell the other AI companies). But surging demand and sustainable business model are two very different things, and the filing doesn't tell us anything about the latter yet.
The timing is interesting too. OpenAI has been talking about going public for months, Sam Altman has been doing the rounds, and then Anthropic just, sort of quietly files first? That feels less like winning a race and more like someone who saw an opening and took it before conditions changed. Smart, probably! But let's not pretend this is about AI supremacy.
This is about securing capital while the getting is good. And right now, with AI still in its hype cycle peak (or near it, who knows), the getting is very good indeed.
The confidential filing period can last a while. Anthropic will go back and forth with the SEC, probably several times. At some point they'll file publicly, we'll get to see the actual numbers, and then we'll have a much better sense of whether that $965 billion valuation makes any sense at all.
My guess, and this is just a guess based on watching a lot of IPOs over the years, is that the public filing will reveal some uncomfortable truths about burn rate and customer concentration. Every AI company right now is dependent on a handful of enterprise customers and a lot of API usage that may or may not stick around. That's not a criticism, it's just the reality of building in a new market.
Will Anthropic's IPO be successful? Probably! The appetite for AI stocks is enormous right now, and Claude has genuine technical credibility that ChatGPT sometimes lacks. But success in an IPO and success as a long-term public company are, again, two very different things.
I've been wrong before (ask me about my 2015 predictions on VR, actually don't), but what I'm seeing here is a company making a rational decision to access public capital while it still can, before the next downturn or the next AI winter or the next whatever. That's not a bad strategy! It's actually pretty smart. But let's not dress it up as some grand victory in the AI wars.
It's just business. The oldest kind of business there is, which is getting other people's money before you run out of your own.
If you want to argue with me about this, my email's on the about page. I check it more than Slack, because I'm a person of a certain age and I have standards.
