AI buildout needs ~$5.5 per GPU-hour revenue, within current pricing: Columbia report
rohanpaul_ai · x · 2026-09-26
A Columbia Business School report, "Financing the AI Buildout," analyzes the central scenario of 182.7 GW of added US data-center capacity between 2025 and 2032. At GB300 NVL72 configurations that implies 1.07M racks / 77M GPUs, requiring $3.725T in mature annual revenue — about $5.5 per installed GPU-hour at full utilization. That sits below the $6–$10+/GPU-hour fetched by current high-end NVIDIA capacity, meaning the buildout's math works without radically higher compute prices.
More from Infra
- NVIDIA at $5.4 trillion is now worth more than the entire UK or French stock market — iamfakhrealam · 2026-09-26
- DeepSeek V4.1 Flash's Engram memory layer trades FFN compute for lookup tables, SemiAnalysis data suggests — teortaxesTex · 2026-09-26
- A Curated Paper List for Learning Distributed LLM Training and Inference — East-Muffin-6472 · 2026-09-26
- vLLM adds Elastic Expert Parallelism: grow/shrink MoE GPU pools under live traffic — PyTorch · 2026-09-26
- AI data center investors now favor real infrastructure over PowerPoint promises — TansuYegen · 2026-09-26
- Running Qwen 27B Q4_K_M on dual RTX 3060 with llama.cpp hits ~44-50 tok/s — jacek2023 · 2026-09-26