Particle.news
Download on the App Store

Goldman Says AI Investment Boom Still Early as Debt-Fueled Risks Draw Warnings

Hyperscaler spending is surging with financing tilting toward debt.

Overview

  • Goldman Sachs compares today’s AI surge to the 1997–98 buildout phase and says the investment cycle has room to run, though strong returns are not guaranteed.
  • Alphabet, Microsoft, Meta and Amazon are set to spend about $375 billion on AI infrastructure and chips this year, rising toward roughly $450 billion next year, according to The Age.
  • Financing is shifting beyond cash flows to bonds, off‑balance‑sheet vehicles and securitised data‑centre debt, with even the largest players tapping credit markets.
  • The Bank of England warns a reset in AI asset values could transmit through multiple channels given uncertain monetisation and bottlenecks in power, data and chips, with AI‑linked stocks now about 44% of the S&P 500.
  • Scale estimates remain vast as McKinsey pegs data‑centre needs at $5.2 trillion by 2030 and Morgan Stanley projects nearly $3 trillion of AI capex in 2025–28 with about half from external capital, while The Age reports some start‑ups like OpenAI have large purchase commitments without secured funding.