Anthropic's Profitability Claim and Its Watchdog Come Under Fire on the Same Day
Anthropic said it's 'highly profitable' if you exclude some of its biggest costs, while critics turned their attention to the funding ties between the company and the safety evaluators meant to hold it accountable.
Anthropic had a strange kind of day: the company claimed profitability and drew accusations of regulatory capture, more or less simultaneously. According to a widely circulated post from @Polymarket, Anthropic says it is 'highly profitable if you take out some of their biggest expenses.' That phrasing did the company no favors. Stripping out your largest costs — which for a frontier lab means compute and training runs — is not a description of profitability so much as a description of a business before its business exists. The framing was picked apart within hours.
The more consequential story running in parallel concerns who evaluates Anthropic's safety work, and who pays them. In a pointed thread, @RWMaloneMD argued that Anthropic 'has built a regulatory capture machine,' singling out METR — a model-evaluation organization — as being 'financially dependent on Anthropic's success.' The claim is that the entities auditing the safety of frontier models are not independent of the labs whose models they audit. If true, that undermines the entire premise of third-party evaluation as a governance mechanism.
Get our free daily newsletter
Get this article free — plus the lead story every day — delivered to your inbox.
Want every article and the full archive? Upgrade anytime.
No spam. Unsubscribe anytime.