The current AI investment landscape represents an earnings bubble rather than a valuation bubble, as the long-term profitability of AI service providers remains uncertain despite massive capital expenditure. While "shovel sellers" like memory chip manufacturers currently report significant earnings, these gains face risks from potential overcapacity and future competition, particularly from Chinese firms that could commoditize AI services. Hyperscalers are aggressively leveraging their balance sheets through debt and lease commitments to fund infrastructure, yet the broader economic impact of a potential AI correction appears manageable. Unlike the TMT cycle, the wealth effect is less pronounced, and non-AI sectors of the market demonstrate resilience. Ultimately, the sustainability of AI-driven returns depends on whether companies can capture value beyond the initial gold rush, as current high capital spending faces a high depreciation hurdle and limited long-term competitive moats.
Sign in to continue reading, translating and more.
Open full episode in Podwise