Ex-Citadel Quant Releases Deep Guide on AI Power Pricing and Data Centers
PandaAshwinee · x · 2026-08-05
Neel Somani, a former quant researcher at Citadel, published a comprehensive primer titled Power 2026, diving deep into electricity pricing and data center development in the AI era.
The article argues that power, not chips, is the true bottleneck for scaling AI compute. Data centers currently account for 5% of US power consumption, with demand doubling every two years, theoretically outpacing total US power generation by the mid-2030s.
The guide is structured into two main parts:
- Part 1: For founders and investors, covering how power plants work, data center development, and commodity pricing fundamentals.
- Part 2: For traders, explaining how to price US power markets.
More from Infra
- Simon Willison Tests Running MiniMax Video Model Locally on Mac — mishig25 · 2026-08-05
- How iota Handles Unreliable GPUs in Distributed Training: A Case Study from Orion 16B Run — markjeffrey · 2026-08-05
- Chrome Extension Uses WebGPU and Gemma for On-Device Text Summarization — jason_mayes · 2026-08-05
- LLMs Master Kernel Writing: Unlocking Complex Optimizations but Making Validation the New Bottleneck — mike64_t · 2026-08-05
- RX 7900 XTX vs RTX 3090 for Local LLMs and Video Generation on Linux — AnimalPuzzleheaded71 · 2026-08-05
- OpenAI Engineer Details WebRTC Improvements Powering GPT-Live — juberti · 2026-08-05