Acemoglu: AI needs $3.7T annual revenue by 2032 to recoup 3.6%-of-GDP investment
ylecun · x · 2026-10-02
Nobel laureate Daron Acemoglu is running a daily series on AI's economics. Citing a paper by Stijn Van Nieuwerburgh, he notes AI investment will average 3.6% of US GDP annually from 2025–2032; at a 10% required return, the industry would need $3.7 trillion in annual revenue by 2032 to recoup the spend — up from roughly $200 billion today, far exceeding 10% of national income. Acemoglu's unasked question: can the AI boom continue without massively increasing inequality?
More from AGI Musings
- Debater cites Second Temple's fall to challenge Scott Alexander on AI doomerism — JoshPurtell · 2026-10-02
- LinkedIn and Twitter's agent access blocks are holding back the agent economy — AccBalanced · 2026-10-02
- AI consciousness discourse returns, prompting philosophy-degree regret and Batman jokes — AndyMasley · 2026-10-02
- "AI welfare is an ideology with thin technical support," researcher fires into consciousness debate — gerardsans · 2026-10-02
- arXiv submissions hit liftoff: real AI-accelerated research or slop flood? — CSProfKGD · 2026-10-02
- 70% of Baidu's mobile search pages have AI content as China closes the commerce loop — Sanbi_Ai · 2026-10-02