The Risk at the End of the Whip | GMO’s Tom Hancock on Finding Conviction Amid the AI Hype

This episode of Excess Returns features GMO’s Tom Hancock on how to think about AI as an investment opportunity and what truly defines “quality” in today’s market. The conversation breaks down the AI value chain, challenges common assumptions about where value will accrue, and ties it all back to building durable portfolios in a rapidly changing technological landscape.Tom walks through his “Hype vs High Conviction” framework, explaining why identifying the right layer of the AI ecosystem may matter more than simply betting on the theme itself, and why balance sheets, durability, and capital allocation remain critical even in the most exciting growth environments.Hype vs High Convictionhttps://www.gmo.com/americas/research-library/hype-vs-high-conviction_insights/Topics Covered:Why AI may be the most important investment decision todayThe four-layer AI stack: applications, LLMs, hyperscalers, and infrastructureWhy investors confuse secular trends with investable opportunitiesFollowing the money through the AI value chainThe hidden risks of investing lower in the stackWhy today’s tech leaders differ from the dot-com eraGrowth vs maintenance capex and what it means for AI economicsWhy software may be more resilient than markets thinkHow GMO defines “quality” and why it matters in volatile marketsPortfolio construction: where GMO is investing (and avoiding) in AITimestamps:00:00 Intro and framing the AI investment debate00:00:55 Tom Hancock background and focus on quality investing00:02:00 What investors are getting wrong about AI00:03:23 Breaking down the four layers of the AI ecosystem00:06:45 Applications vs infrastructure: where value may accrue00:08:45 Why predicting AI winners is still difficult00:11:00 Following the cash flows through the AI stack00:13:00 Why AI funding is more stable than past tech bubbles00:16:00 Big Tech strategy differences and capital allocation decisions00:17:34 Are today’s tech companies higher quality than in 1999?00:19:00 Growth vs maintenance capex and implications for Nvidia and others00:22:00 Depreciation, chip lifecycles, and hidden risks in capex assumptions00:24:00 Capital intensity vs quality: when heavy investment is a feature00:27:00 Why incumbents may benefit most from AI00:28:30 Risks in the LLM layer and potential commoditization00:30:10 Software disruption fears: overdone or justified?00:34:06 Defining “quality” in investing00:36:00 Balance sheets vs return on capital00:38:32 Why GMO sold Oracle and the risks of leverage00:40:18 What happens if AI spending slows down00:41:35 Where the biggest risks are in the AI stack00:44:26 Where GMO is positioned vs the S&P 50000:48:00 How new ideas enter a quality portfolio00:51:00 Sell discipline and portfolio turnover00:53:00 International vs US quality investing