Nvidia and its peers—Meta, Apple, Microsoft, Alphabet, Amazon and Tesla—once dominated the US market. In June they lost more than $2tn of market value, a fall driven by fears that the colossal AI‑infrastructure outlays—over $1tr in 2025‑26—will not pay off. Rising component costs and debt‑dependency have squeezed margins, turning the once‑magnetic group into a fragmented set of stocks.

The shift from value to momentum investing has left the Mag 7 as a rallying point for index‑fund flows, a practice that can distort prices. Citigroup analysts now declare the construct dead for assessing growth dynamics, while hedge funds take long positions in semiconductors and short hyperscalers, amplifying volatility.

Benchmarking pressures, as studied by Paul Woolley and Dimitri Vayanos, force managers to trade into overpriced names to stay on track, creating macro‑economic misallocations. The debate now centers on whether we are witnessing a Mag 7‑induced bubble, a scenario that would require a compelling story, leverage, and fear of missing out.

AI’s potential to transform knowledge creation and labor is unprecedented, yet history shows that real benefits often follow a period of disappointment after a bubble bursts.