What happened
Stanford's One Hundred Year Study on Artificial Intelligence released its third five-yearly report, 'AI Goes Mainstream: The Scaling Era and the Rise of Generative AI,' on 10 November 2026. The international panel, chaired by Oxford professor Michael Wooldridge, concludes the scaling paradigm — more compute and more data — drove generative AI's advances but is 'running into data and compute constraints, shifting the frontier toward reasoning and inference-time compute.' It documents AI compressing scientific research timelines 'from years to months or weeks,' notes frontier AI development is 'concentrated among a small number of U.S. and Chinese actors,' and flags cybersecurity and military AI as major emerging risk areas. It delivers a central qualification to the investment boom: 'AI's economic impact was not yet evident in aggregate macroeconomic data or companies' returns on investment.' The report also finds the public 'has had little say in how AI is developed and deployed.'
Why it matters
This is the flagship longitudinal benchmark on where AI is and where it is headed — the framing executives use to set strategy, and a caution on the widening gap between AI investment and demonstrated economic returns.
Action needed
Brief the board on the 12 key takeaways, especially the scaling limits, the concentration of frontier AI, and the gap between the AI investment surge and demonstrated macro/company returns.