Enterprise AI fails when facilities aren't included in planning

Aaron Dudley argues that organizations must include facilities teams in AI planning to avoid costly infrastructure retrofits.

• Main announcement: Aaron Dudley recommends starting AI infrastructure planning with facilities (not procurement) to avoid late-stage, project-ending costs — exemplified by a $2.8 million quote to pipe water into a building never designed for GPU workloads; he also highlights the planning question of “If we had a $1M budget to prepare for AI, where would you invest first?” (Article date: April 20, 2026).
• Background and details: He outlines a standardized assessment approach (verify transformer capacity, structural limits, water availability) and the use of digital modeling tools (e.g., ETAP) to simulate power, cooling, and structural scenarios before procurement; this reframes risk from the physical world to digital tests and shifts responsibility to include facilities teams early.