AI capital expenditure is a boom, not a bubble — Jared Gross

Jared Gross of J.P. Morgan Asset Management offers a TLDR of his 55-page report, outlining two paths taken by past capital expenditure booms. He says AI is following a highly competitive path like railroads, rather than the monopoly path of AT&T or Standard Oil, and that destruction will intensify after capital inflows peak. He says a bubble implies an illusion, but AI is a real industry and an economic engine that will transform markets, the economy, and society. While overvalued companies and bankruptcies will emerge, he emphasizes that this is a capital expenditure boom of unprecedented scale and speed. From an asset allocator's perspective, he calls for diversification across themes as existing diversification tools are swept up in the wave. He explains that the current picks-and-shovels phase will be followed by a process of identifying the AI winners among end users and a global spread of productivity gains. He says it is difficult to imagine AI evolving without losers: the key issues are suppliers' assessments of excessive capital expenditure and AI operators' ability to generate sustainable revenue. He says unit computing costs are falling, with low-cost Chinese models setting a price ceiling. He explains that hyperscaler debt starts from a position of strong earnings and finances, limiting credit risk and adding diversification to the investment-grade market. Overall spreads are not concerning, but individual securities with widened spreads require discrimination through active management. All directional outlooks are the speaker's conditional judgments, not definitive predictions of short-term prices.