AI Bubble Risks: Energy Costs and Policy Uncertainty

Policy risk is another factor. Hayes cited US restrictions on Anthropic's Mythos and Fable models, which limited access to American users. The move, he said, shows how policy decisions can disrupt service without warning. For foreign users, the risk of losing access to premium AI tools creates uncertainty that is not reflected in current valuations, Hayes argued. Businesses built around continuous AI service may need contingency plans.

Open-source alternatives are also eroding the pricing power of US AI companies. Hayes said Chinese-developed open-source models often cost roughly one-tenth of comparable US offerings while delivering similar performance. Users also gain greater control over their data, he said. Cost and control together make these alternatives increasingly attractive, Hayes concluded. The shift threatens the high margins that justify current AI valuations.

Hayes did not offer a timeline for a correction but said the three risks are interconnected and could compound. The market's assumption that AI growth will continue uninterrupted may be tested, he said.