AI Eats the World? A Reality Check with Benedict Evans

The AI transition mirrors previous platform shifts like the internet and mobile, characterized by initial uncertainty, massive infrastructure investment, and a search for sustainable product-market fit. While AI has demonstrated immediate utility in coding—evidenced by a dramatic surge in run rates—broader daily consumer and enterprise adoption remains elusive. Foundation models likely function as commodity infrastructure, forcing value to migrate toward specialized applications that solve specific industry problems. The current surge in capital expenditure represents a period of extreme market disequilibrium, as companies scramble to participate in a fundamental shift without clear pricing power or established business models. Ultimately, AI will follow the path of past technologies, evolving from a disruptive, expensive novelty into an invisible, ubiquitous utility that reshapes organizational workflows and creates entirely new economic categories.