Is It a Bubble?

Howard Marks of Oaktree Capital analyzes the AI investment boom through the lens of bubble history, distinguishing between "mean-reversion bubbles" (purely financial manias) and "inflection bubbles" (driven by genuine technological progress). While AI exhibits classic bubble characteristics—exuberant speculation, FOMO, circular deals, and $1B+ seed rounds—Marks notes key differences from the dot-com era: existing products with exploding demand, established profitable players, and lower valuation multiples. He highlights three major risks: (1) massive debt financing for speculative infrastructure with uncertain payoffs; (2) unpredictable demand and competitive dynamics that could render today's leaders obsolete; (3) the possibility of overcapacity in data centers. Marks concludes that while AI's transformational potential is real, investors face genuine risk of losses if enthusiasm proves excessive, and recommends moderate, selective positioning rather than all-in or all-out approaches.