Economists are examining signs of a potential AI-driven market bubble, using historical patterns like high valuations, volatility, and new stock surges as clues. A 2019 research framework challenges the idea that bubbles are unpredictable, though its accuracy remains limited. While current markets show some red flags — especially in tech giants like NVIDIA — the lack of widespread new stock issuance weakens the bubble case. Unlike past crises, an AI bust may pose less systemic risk due to minimal bank exposure, but could still waste resources, such as underused data centers. Yet, speculation isn’t purely destructive: history shows that even failed booms, like the dot-com era, can leave behind valuable infrastructure and innovation. This raises a deeper question — should society fear all bubbles, or might some overinvestment actually accelerate progress?