Chris Wood’s big warning: The specific risk that will finally trigger the end of AI trade

Chris Wood warns that the AI trade is most likely to end not from a traditional chip oversupply but from malinvestment — hyperscalers and AI labs failing to earn adequate returns on their massive capex. He highlights TSMC's raised 2026 capex guidance of about 56 billion US dollars and Taiwan's real GDP growth of 14.55 percent year-on-year in Q1 2026 as evidence of the dramatic build-out, while flagging circular financing structures like Nvidia funding OpenAI's chip purchases as a key vulnerability. Wood also points to rapid commoditisation of AI models, particularly cheaper Chinese alternatives on platforms like OpenRouter, as adding pressure on premium AI provider economics.