Recent OpenAI revenue signals falling short of prior expectations have heightened trader caution around AI infrastructure spending, with hyperscalers guiding toward hundreds of billions in 2026 capex amid token price collapses and intensifying competition from cheaper Chinese open-weight models. Warnings from figures like Michael Burry and Ray Dalio, alongside BIS notes on exuberance risks, underscore concerns over debt-fueled data center builds, energy costs, and whether large language model adoption will justify valuations before 2027. Key swing factors include upcoming earnings from labs like Anthropic, hyperscaler capex updates, and any regulatory or supply-chain shifts that could alter the competitive landscape between U.S. frontier models and global alternatives.