Is AI a Bubble? [Exponential View] (www.exponentialview.co)

An Exponential View essay lays out a practical, data-driven framework to answer whether today’s AI surge is a bubble. Rather than declaring judgement, the author builds a five-gauge dashboard—economic strain, industry strain, revenue growth, valuation heat and funding quality—benchmarked against past tech bubbles. The piece defines a bubble operationally (a ≥50% equity drawdown sustained for ≥5 years plus a ~50% fall in productive capital), notes that bubbles are only obvious in hindsight, and argues we should instead compare current capex and revenues to historical episodes. The analysis and underlying data will be published for subscribers, and a limited number of advisory sessions are being offered to investors and executives. Key technical signals and numbers show why this matters to AI/ML practitioners and financiers: hyperscalers more than doubled data‑center capex since late 2022, Morgan Stanley projects ~$3T in AI infrastructure spend by 2029, McKinsey estimates ~156 GW of new power capacity and $5–7.9T of capex to meet AI demand by 2030, and global data‑center spend may hit ~$370B in 2025 (~0.9% of US GDP if U.S.-weighted). The framework helps distinguish a durable “boom” (where fundamentals catch up) from a speculative bubble (where capital and narratives outpace real productivity), giving teams a structured way to stress-test strategies amid massive, economy‑bending investment.