Anthropic IPO Filing

Anthropic just confidentially filed its S-1. I had to read the revenue line twice.

$47B run rate. Up from $10B a year ago. Faster than Snowflake. Faster than ServiceNow. Faster than Salesforce in its hottest stretch.

But the headline number isn't what makes this filing matter. For two years, the AI conversation has lived on model benchmarks, capex commitments, and private round markups — none of which translate cleanly to a public market. An S-1 does. It forces gross margin, churn, customer concentration, and compute cost into daylight for the first time at this scale.

Claude Code is the wedge that got them here. Developers picked it. Enterprises followed. Ramp's data showed Anthropic overtaking OpenAI in paid business adoption last month. Now the public market gets to vote — and likely before OpenAI does, with a Q4 listing reportedly targeted for October.

Three things I'll be watching when the public S-1 drops: how concentrated enterprise revenue is across the top 10 customers, what percentage comes from Claude Code versus general API usage, and the actual cost of compute as a percentage of revenue — the number every founder building on top of frontier models needs to know.

The IPO race ahead of OpenAI is the story everyone is leading with. The unit economics inside the filing are the story that will actually matter — and the reality check the entire AI thesis has been waiting for.

What's the single line item you'd most want to see on Anthropic's S-1?

https://www.cnbc.com/2026/06/01/anthropic-ipo-s1-prospectus.html