rvachev.orgEN / RU / 🤖
← Back to essays
· Essay · 1 min

Anthropic Updates S-1 Before IPO

Anthropic updated its S-1 before the IPO. An interesting phrase in the risk section suggests AI may soon achieve significant acceleration.

⚡️ Anthropic has updated its S-1 before the IPO. The most interesting part is a phrase in the risk section.

Literally: "our internal estimates and our expectations regarding future models, rather than our current models, indicate that we may imminently meet our Responsible Scaling Policy's definition of 'dramatically accelerated progress attributable to AI'."

This means that according to the company's internal assessments, its next models may soon cross the threshold where AI significantly accelerates its own development. This is what is usually called recursive self-improvement. At Anthropic, they call it "closing the loop": each version of Claude helps build the next one. The current models are not formally recognized as having crossed the threshold, but the company itself writes that significant acceleration may only be noticed after it has already occurred.

The scale of this cycle is evident from the numbers. The cost of internal use of Claude at list prices has increased more than 250 times from January to September 2026, and since April it has exceeded all salary expenses (excluding stock). At the same time, Anthropic specifically states that it does not plan to allow Claude to autonomously build the next generations of models without human oversight and is considering updating the RSP to ensure humans remain in the loop.

The finances are also impressive:
- Revenue for the first 9 months of 2026 is $33.2 billion compared to $2.5 billion a year earlier, with $4.6 billion for the entire 2025
- Customers paying more than $100k a year have increased to over 11,900, up from 950 a year ago
- Operating loss for 9 months has reduced to $1.3 billion from $5.2 billion
- Total commitments for compute and infrastructure are about $895 billion until 2045, with $447.5 billion by the end of 2030

What do you think, is Anthropic really on the verge of RSI or is this just a buildup before the IPO?

https://www.sec.gov/Archives/edgar/data/2130357/000162828026068676/anthropicpbc-sx1ax1.htm

https://finance.yahoo.com/technology/article/anthropic-reveals-amended-s-1-outlining-huge-spending-commitments-rising-sales-153635825.html

#anthropic@rvnikita_blog #claude@rvnikita_blog #ipo@rvnikita_blog #agi@rvnikita_blog #safety@rvnikita_blog

Anthropic Updates S-1 Before IPO — illustration

Newsletter

Get new posts by email

A weekly digest of what I write about AI, software and products. Unsubscribe anytime.