What changes when the company behind your coding assistant has to answer to the stock market every three months? Starting today, we get to find out.
Anthropic, the company behind the Claude models, priced its initial public offering on Wednesday, November 18, at $83 a share. That is above the $70 to $78 range it had set on November 10. The shares start trading today, Thursday, November 19, on the Nasdaq under the ticker ANT. Reuters reports it is the largest stock market debut on record, and the first by a major frontier AI lab.
The numbers, in plain words
The offering covers 980 million shares, more than the 825 million first planned, raising about $81.3 billion from public investors. A separate $26 billion private placement sits alongside it, bringing the total deal to about $107.3 billion.
You will see two different values for the company, and both are honest. Counting only the shares that exist today, Anthropic is worth about $1.1 trillion. Counting also the stock options and restricted shares that employees hold, the fully diluted value is about $1.5 trillion. Think of it like counting a house with the furniture included or not. Same house, different count.
The growth behind those numbers is real. Anthropic's revenue run rate reached $81 billion a year as of October 31, and 11,900 customers were each paying at least $100,000 a year as of September 30, up from more than 1,500 at the end of last year. Reuters notes that Claude Code, its coding agent, has been a big part of that growth, which will surprise nobody who has watched developers adopt it.
The bill is just as big
Here is the other side. Anthropic spent nearly $9 billion on computing capacity in the first half of 2026, about four times what it spent a year earlier, plus another $5 billion in the third quarter. It has committed to nearly $900 billion in future computing spending with partners including Amazon and Google. In the third quarter it posted an operating loss of $567 million, even as revenue hit $16.96 billion.
Photo by Kevin Ache on Unsplash
That is the tension every frontier lab lives with. The models are popular, the revenue is enormous, and the compute bill is more enormous. Going public does not remove that tension. It just makes it visible, in quarterly reports anyone can read.
What this means for Claude, and for us
Until now, Anthropic could set its prices and its pace in private. As a public company, it will report results every quarter, and investors will ask the obvious question: when does all this spending turn into profit? That pressure could push in two directions. It might mean higher prices or tighter limits for heavy users. Or it might mean the opposite: a well funded Anthropic can afford to keep Claude competitive while smaller rivals struggle to pay their compute bills.
There is also the safety question. Anthropic built its name on being careful, and its own filing warns that more advanced models could cause harm. CEO Dario Amodei has argued the industry should slow down. Public shareholders, though, usually reward speed and growth. How those two instincts live together inside one company is something we will now watch in public, quarter by quarter.
Reuters also notes that OpenAI, Anthropic's closest rival, is expected to pursue its own listing as early as next year. So this is probably not a one off event. The closed model labs, the ones whose weights we cannot download or inspect, are becoming public companies with public numbers.
I find that genuinely interesting. For years, the AI race has been argued with benchmarks and vibes. Now at least one of the biggest labs has to show its homework: real revenue, real costs, real losses, four times a year.
So here is my question for you: when the labs behind our tools have to answer to shareholders, does that make the tools better, or just more expensive?
Top comments (0)