AI Valuations: Former OpenAI Expert Calls Private Company Figures 'Totally Made Up'

Miles Brundage, a prominent AI policy researcher and former head of OpenAI's AGI Readiness team, has critically assessed the current state of private AI company valuations, stating they are "essentially totally made up" and "almost certainly massively off in one direction or the other in every case." His remarks, shared via a recent tweet, highlight a growing concern within the rapidly expanding artificial intelligence sector.

The challenges in accurately valuing private AI firms stem from several unique factors. Unlike traditional companies, AI startups often possess intangible assets such as proprietary datasets, unique model architectures, and highly specialized technical teams, which are difficult to quantify using conventional financial metrics. These companies frequently operate with high burn rates due to significant investments in compute resources and top-tier talent, often before generating substantial revenue.

This creates an "AI valuation paradox," where companies with minimal current earnings command astronomical valuations based primarily on speculative future potential. Investors are increasingly betting on technological breakthroughs and market dominance rather than established cash flows. Recent examples include Anthropic's valuation climbing to $183 billion and OpenAI reportedly seeking a $500 billion valuation, demonstrating the intense investor interest and the perceived strategic importance of AI.

Despite the enthusiasm, experts warn of the risks associated with this "gold rush" for AI investments. The lack of standardized valuation frameworks and direct market comparables for diverse AI business models contributes to uncertainty. This environment raises concerns about potential overvaluation and the risk of an "AI bubble," necessitating rigorous due diligence from private equity firms to distinguish genuine value from inflated claims.