The Reasoning Show Understanding NeoClouds with Crusoe
34 snips
Mar 8, 2026 Erwan Menard, SVP of product management at Crusoe who built their NeoCloud strategy, explains a purpose-built AI cloud. He discusses why a new AI-dedicated cloud category matters. He outlines linking energy, modular data centers, and cloud services. He covers challenges of synchronizing energy deals, builds, and rapid software change, plus sovereign and private AI needs.
AI Snips
Chapters
Transcript
Episode notes
NeoClouds Are Born From Demand And Physics
- NeoClouds arise from converging massive AI demand and a need to reinvent the entire infrastructure stack.
- Erwan Menard notes GPU racks jumped from ~15 kW to ~115 kW, forcing new data center physics and purpose-built clouds.
Crusoe Sells From Electrons To Tokens
- Crusoe ties energy sourcing, data center development, and cloud software together to deliver AI services from "electrons to tokens."
- Menard stresses cost and sustainability require connecting energy deals, facility builds, and GPU/software stacks end-to-end.
Synchronize Energy Builds Software And Data Centers
- The hardest coordination challenge is synchronizing three streams that operate at different paces: multi-year energy deals, quarters-long site builds, and week-long software sprints.
- Menard cites a 1.2 GW Adeline campus built into production in 11 months as evidence timelines are compressing.