Paper: AI buildout to cost $10.3T over 8 years, needs $3.7T annual revenue by 2032
AccBalanced · x · 2026-09-26
A new BPEA paper by Stijn Van Nieuwerburgh estimates the AI infrastructure buildout will cost $10.3 trillion over the next 8 years. To deliver a 10% return, investors would need $3.7 trillion in annual revenues by 2032 — roughly 9% of US GDP.
The thread also touches on datacenter financing: why would Meta pay more in interest just to nominally remain an "asset-light" business? The linked writeup is called one of the clearest explanations of datacenter finance.
More from Venture
- Similarweb: Claude logs record 1.253 billion website visits in October — gaganghotra_ · 2026-11-11
- a16z Podcast: Mike Volpi on What the 1990s Internet Boom Teaches Us About the AI Buildout — a16z Podcast · 2026-11-11
- Alibaba files to list chip-design unit T-Head in Hong Kong to fund AI push — pstAsiatech · 2026-11-11
- First traffic snapshot of Muse, Instinct, Hark and Fo: 60M visits in October, Muse leads with 45M — agentaya-ltd · 2026-11-11
- AI killed SEO writing moats; first-hand data is the real edge — sven_ai · 2026-11-11
- Investor makes the case: HBM at 3x forward earnings with 60% growth is undervalued — firstadopter · 2026-11-11