Debt Is Fueling the Next Wave of the AI Boom (www.wsj.com)

What’s changed: after a few years of AI growth fueled largely by big-tech cash from advertising and cloud services, an increasing amount of investment is now being supported by debt — a dynamic that is reviving comparisons to the dot‑com bubble. Early-stage AI scaling didn’t look bubble‑like because companies were spending free cash on compute and talent; today, cheaper borrowing and leveraged capital are underwriting an aggressive build‑out of datacenters, GPUs, and acquisitions. Why it matters for AI/ML: debt accelerates model scale and productization by supplying large, fast pools of capital for expensive infrastructure (GPU farms, custom accelerators, networking) and M&A, enabling faster iteration on large models and production deployments. But it also raises systemic risks: pressure for near‑term returns can skew R&D toward revenue‑driven features over foundational science, increase the chance of overcapacity or stranded hardware if funding tightens, and intensify consolidation pressures on startups. For researchers and engineers, the shift means more compute availability in the short term but also greater volatility and potential shifts in hiring, open research priorities, and long‑term investment if macro conditions or credit markets sour.