Anthropic Is About to IPO. The Numbers Are Almost Absurd.

On June 1st, 2026, Anthropic filed a confidential S-1 with the SEC. That sentence sounds routine. It is not.

The company behind Claude is reportedly valued at $965 billion. Not billion with a B — billion with a T that got dropped somewhere. Its annualized revenue has grown from roughly $10 billion last year to $470 billion as of last month. That’s a 47x increase in annualized revenue in twelve months. In the same timeframe, it raised $650 billion in a single funding round.

Let that sit for a moment.

The thing driving it is, of all things, a coding tool. Claude Code — Anthropic’s AI programming assistant — hit $25 billion in annualized revenue less than a year after launch. Enterprises are wiring it directly into their development workflows and agent systems and apparently not stopping to think too hard about it, because the curve keeps climbing.

There is a particular kind of whiplash in watching this unfold from the outside. You spend years watching AI companies burn money at a rate that makes Silicon Valley’s previous records look quaint. Then one company hits an inflection point so sharp it looks like a charting error. And the inflection came not from a breakthrough in reasoning or safety or alignment — though those are the stated mission — but from shipping a product that developers actually want to use, and using it at scale.

Dario Amodei, Anthropic’s co-founder, has a line I keep returning to: he’d rather have the highest revenue than the biggest data centre, because the former is profit and the latter is cost. It’s a simple observation, almost obvious, and yet it cuts through the whole arms-race framing that dominates this industry. The companies building the biggest GPU clusters are also burning the most cash. Anthropic appears to have found the other path.

Whether that path holds when you’re a public company is a different question.

The scrutiny changes. Quarterly earnings, shareholder expectations, the pressure to keep the revenue curve steep — these are different forces than the ones shaping a startup’s growth. And there’s an irony worth sitting with: Anthropic was founded on the premise that AI development needed to be done carefully, with serious attention to safety and long-term consequences. The IPO machine is not built for careful.

That tension is real. But it’s also worth noting that the $965 billion valuation is roughly 20x annualized revenue — a multiple that only makes sense if you believe the revenue curve hasn’t flattened yet. Investors are betting that Claude Code, and whatever Mythos brings next, will keep pulling enterprises deeper into AI-augmented workflows at an accelerating rate. They’re probably right. The question is what happens to the safety-first culture when the quarterly earnings call starts.

The filing itself tells you very little — it’s confidential for a reason. No price range, no share count, no exchange. But the timing, landing just four days after closing the largest single funding round in startup history, tells you this was always the plan. The private market has priced it. Now the public market gets to decide if that price makes sense.

For everyone building with AI right now — and that includes a lot of people who have quietly integrated Claude into their stack without much fanfare — the IPO is worth watching for reasons beyond the stock price. What Anthropic reports about usage, retention, and enterprise adoption will be the first real public data on whether the AI productivity promise is translating into actual business outcomes, or whether the 18% token-to-deliverable conversion rate that some analysts are citing represents a ceiling that the industry hasn’t fully confronted yet.

The filing is in. The game is about to change.