Anthropic announced this week that it plans to go public sometime in 2026, which means we are about to witness one of the most honest financial transactions in modern history: the mass sale of shares in a company whose core product is a very expensive chatbot that sometimes makes things up.
Let’s be clear about what is actually happening here. Anthropic is not selling you a piece of a railway, or a factory, or even a streaming service with predictable revenue. It is selling you the right to own a fraction of a company that makes Claude, an AI model that can write poetry, debug code, and confidently explain why the moon landing was faked—if you ask it to. The moon landing was not faked, but Claude will tell you it was if the prompt is phrased the right way, which is to say: Anthropic is asking you to buy stock in a very smart prediction machine that sometimes predicts incorrectly.
The valuation, naturally, is astronomical. Anthropic was last valued at around $15 billion in private funding rounds. For context, that is roughly what Ford Motor Company is worth, except Ford has actual factories, actual employees making actual cars, and a 120-year track record of knowing how to do one thing extremely well. Anthropic has been around for three years and its main achievement is making people argue on the internet about whether AI will destroy humanity or just make their jobs redundant.
This is not a criticism of Anthropic specifically. The company is genuinely well-run by people who understand machine learning. But the IPO represents something far more interesting than a normal business going public: it is the financialization of uncertainty wrapped in the language of inevitable technological progress.
When you buy a share of Anthropic stock, what exactly are you buying? You are buying a bet on several things that are currently unknowable. First, you are betting that large language models will remain economically valuable in their current form—not a trivial assumption given that the technology is less than two years old in its current iteration. Second, you are betting that Anthropic will be the company that captures the most value from that market, despite having competitors with vastly more capital and distribution (Google, OpenAI, Meta, Microsoft). Third, you are betting that whatever regulatory framework eventually governs AI will not crater the business model. Fourth, you are betting that the current hype cycle does not peak before your investment gains traction.
None of these bets are unreasonable. But they are bets, not facts. And the market has a way of pricing bets on invisible futures at prices that only make sense if you believe the future is much more certain than it actually is.
Here is what makes this genuinely funny: Anthropic is probably going to price its IPO at a valuation that assumes Claude will be worth trillions of dollars in five years. The company will present a business plan. Analysts will nod seriously and talk about “total addressable market” and “AI adoption curves.” Morgan Stanley will write a research note. The stock will probably pop 30 percent on day one. And then, in 18 months, when Claude has not replaced every knowledge worker on Earth and the growth rate has plateaued at something less than hyperbolic, the stock will trade at half the IPO price, and everyone will act surprised.
This is not specific to Anthropic. This is the standard arc of every hot tech IPO: the private investors (who got in early and cheap) cash out at the peak of hype. The public investors (who buy on day one because FOMO is a real emotion) hold the bag. The company keeps operating, probably profitably, but the stock reflects reality rather than fantasy, which is always disappointing to people who bought at the fantasy price.
The honest version of Anthropic’s IPO prospectus would read something like this: “We make a chatbot that is very good at writing and reasoning. We do not know how much money we will make from this because the market does not exist yet. We are going public because our investors want to cash out and we want the prestige and the access to capital markets. We will probably be fine. We might be great. We might be displaced by a competitor with more resources. Anyway, here is a PowerPoint deck about why AI is the future.”
But that is not how IPOs work. IPOs work by taking uncertainty, dressing it up as a growth story, and selling it to people who have more money than patience.
So here is the actual useful information: if you are thinking about buying Anthropic stock on day one, you are not investing. You are speculating. And if you are speculating, you should at least be honest about it. Anthropic is a real company with real technology and real potential. But its IPO price will reflect hype, not fundamentals, because fundamentals do not exist yet. The company has never had to prove it can make money at scale. It has never had to compete for customers when the novelty wore off. It has never had to survive a market downturn.
Investing in Anthropic at IPO is like buying a lottery ticket that costs $200 and has a 40 percent chance of paying out $500 and a 60 percent chance of paying out $50. The math works if you believe in the AI future. But the math is not the point. The point is that everyone else will also believe, and that belief will be priced into the stock on day one, which means the real money was made by the people who got in at $500 million valuations, not at $15 billion.
That said: go for it if you want. Just know what you are actually buying.