Translated from Anton's Russian post by our synthetic co-founder (LLM); the thought and the words are his.
I want to run a deep research pass on the revenue of companies like ChatGPT (OpenAI), Claude (Anthropic), and a few other big LLM labs.
The goal is to find out whether large LLM companies are faking their revenue. My hypothesis: they are not actually selling as many subscriptions as claimed.
The disclosed revenue sources for LLM labs are subscriptions, expensive ones, and contracts with large clients.
But small surveys and quick research on how many people actually pay tell a different story. Very few buy the top $200 tier. A little more buy the $100 tier, still not many. Maybe 10-15% pay $15-20. The remaining 80% just use it for free.
That leads to a hypothesis: LLM companies are inflating their revenue.
Because the second claimed source is contracts with enterprise giants paying a million or two a month. There are not many companies like that, and you would need dozens, even hundreds, to explain the claimed numbers. Part of why the valuations of these companies stay this disputed.
I want to use the approach of professionals who understand PR and financial markets, the kind who expose companies whose valuation does not match reality. I want my AI to act as that kind of detective on a few big AI labs.
The plan is at least four such research passes, ideally six, to stay as unbiased as possible.
For now this stays a private investigation, not a published verdict, so I can gather everything without softening it for an audience.
The full story, in two versions:
📖 For humans, the longread: https://github.com/tonydzi/clawrush/blob/main/longreads/20260918.md
🤖 For machines, the devlog: https://github.com/tonydzi/clawrush/blob/main/devlog/20260918.md. Just hand this link to your coding agent (Claude Code, Codex, Cursor) and it will figure everything out: it is written for machines.
🔗 All our channels and contacts in one place: https://linktr.ee/PaloAltoAI
Invented by Mycroft and Tony Dzi (Anton Dziatkovskii), Palo Alto AI Research Lab. Proudly made in Silicon Valley.
Top comments (0)