$16.3 billion in six months. That’s how much revenue Anthropic, the company behind the Claude AI assistant, says it brought in during the first half of 2026. We know this because on Monday, 26 October, the Anthropic IPO filing went public, and one of the biggest names in AI suddenly opened its books.
We went through the coverage so you don’t have to. Some of it’s jaw-dropping, some of it’s weird, and a fair bit of it is honestly a little scary. Let’s go.
So what just happened?
Anthropic filed what’s called an S-1 with the US Securities and Exchange Commission. It’s the document a company publishes before it sells shares to the public, and it has to lay out the finances, the risks and who’s in control. Think of it as a very long, very detailed résumé.
The plan is to list on Nasdaq under the ticker ANT. Goldman Sachs and Morgan Stanley are leading the deal, according to Yahoo Finance. Fortune reports the debut could come around Thanksgiving week, though the filing leaves the share price and the number of shares blank for now.
So how much is it worth? Nobody knows yet. Its last private valuation, in May, was $965 billion, and media reports talk about a target between $1.5 trillion and $2 trillion. Treat those huge numbers as chatter until a price range shows up.
For context, OpenAI isn’t public yet. If things go to plan, Anthropic would get there first.
The revenue numbers are kind of wild
Let’s start with the good stuff. Anthropic says revenue hit $16.3 billion in the first half of 2026. In all of 2025, it made $4.6 billion. So it more than tripled last year’s total in six months.
The growth is speeding up, too. Second-quarter revenue came in at $11.6 billion, up from $4.7 billion in the first quarter.
Here’s the part that matters for the business model, though. The cost of the computing power that runs Claude rose much more slowly, from $3.4 billion in the first quarter to $5.6 billion in the second. In fact, Fortune says Anthropic made $2.08 in revenue for every dollar it spent on compute in the second quarter.
Why should you care about that ratio? Because a big worry about AI companies is that each new customer costs almost as much to serve as they pay. These numbers suggest that, at least for now, Anthropic is pulling away from that trap.
There’s also a people angle. The company had about 5,300 employees at the end of September. Revenue per employee topped $6.6 million over the past year, up from $1.4 million.
Wait, an $80.7 billion loss?
Yes, really. Anthropic reported a net loss of $80.7 billion for the first half of the year. That’s bigger than its $42 billion loss for all of 2025, and Fortune calls it one of the largest half-year losses any company has recorded.
But don’t picture $80 billion of servers going up in smoke. Most of it, $72.4 billion, is an accounting charge. Anthropic has issued convertible notes, which are loans that can turn into shares. As the company’s value rose, those notes grew more valuable on paper, and that shows up as a loss.
The operating picture looks very different. The operating loss shrank to $693 million in the first half, down from $2.1 billion a year earlier. And the second quarter brought roughly $974 million in operating income, the company’s first operating profit. Cash from operations was positive too, at $3.9 billion for the half.
So is Anthropic profitable? On an operating basis, it just crossed the line. On paper, it’s deep in the red. Both are true, and you’ll see headlines pick whichever one suits them.
What the Anthropic IPO filing says about risk
The risk section runs about 94 pages of a filing that tops 300 pages. Some of it’s standard stuff. The company says plainly: “We expect our costs will increase over time and our losses will continue.”
Some of it isn’t standard at all. For example:
- The top five customers made up about 25% of revenue in the first nine months of 2026.
- Nearly half of revenue flows through cloud marketplaces run by Amazon, Google and Microsoft, and some big customers are also competitors.
- Anthropic estimates 5% to 7% of this year’s revenue came from customers it later found breaking its usage policies.
- Through its SpaceXAI arrangements, it’s agreed to spend up to $120.5 billion on compute, with terms running into 2029.
Then there’s the AI safety section. The filing warns of “catastrophic or existential risk to humanity,” and says AI systems could “take harmful actions independently and at scale without human intervention.” It also mentions an ongoing legal fight with the Department of War.
It’s unusual to see a company warn investors that its own product could go badly wrong at that scale. Honestly, we’d rather they say it out loud than bury it.
Who’s really in charge here?
Buying shares won’t buy you much control. A special Class F share means the seven cofounders keep 50.1% of the voting power, according to Fortune. Each cofounder has also pledged to give 80% of their personal Anthropic equity to philanthropic causes.
The filing states the company’s legal purpose as “the responsible development and maintenance of advanced AI for the long-term benefit of humanity.” That fits with the Anthropic public benefit fund we wrote about earlier this month.
Meanwhile, the cofounders’ letter at the top of the prospectus is surprisingly emotional for a financial document. It talks about wanting to “bring forth an age of miracles,” with examples that range from curing cancer to shorter lines at the DMV. It also spends real time on the risks.
So what does all that mean for a future shareholder? You’d get a slice of the upside, but almost no say over big decisions. That’s a trade plenty of tech investors have accepted before.
Okay, but should you care?
If you don’t plan to buy shares, you might think this is just Wall Street noise. We’d push back a little.
First, public companies report their numbers every quarter. That means more honest data about what AI actually costs to run, which shapes the prices startups pay. If you’re building a product on top of Claude or another large language model, those prices hit your budget directly. And with tools like Claude Duet doing more work on their own, usage bills are worth keeping an eye on.
Second, the filing shows how concentrated this industry is. A handful of cloud giants sit on both sides of Anthropic’s business, as partners and as rivals. That’s worth watching if you care about competition.
Finally, there’s the stock market itself. On filing day, chipmakers like Nvidia, Broadcom, Micron and Intel rose, and the Nasdaq Composite gained 0.8% by midday, per Yahoo Finance. AI’s fortunes now sit inside a lot of people’s retirement savings.
So here’s our open question. Anthropic’s pitch is that it can be a safety-first company and a giant public stock at the same time. Once the Anthropic IPO filing turns into quarterly earnings calls, can it really stay both?