The evolution of AI capacity planning

Joel Morris, Inference Compute Planning Lead, OpenAI joins SDxCentral’s Kat Sullivan to discuss how AI capacity planning is evolving from a traditional compute-focused exercise into a core intelligence and business strategy. Key discussion points include:

  • Why traditional cloud and SaaS capacity planning models break down for AI workloads
  • How GPU capacity has evolved into a product, business, and intelligence-planning challenge
  • What a “North Star” view of AI infrastructure means, and how modern AI capacity planning stacks support it
  • How automation and solver-based allocation are reshaping planning, prioritization, and operations across AI organizations