What does it take to build an IT infrastructure that’s ready for AI? Most enterprise IT environments were designed to run ERP, CRM and databases. The predate generative AI entirely. In this episode, Dell’s Global CTO and Chief AI Officer John Roese breaks down exactly how AI infrastructure differs from traditional IT — from GPU-based compute and purpose-built AI factories to the new knowledge layer that sits above your systems of record. If you’re making architectural decisions today that will shape your AI capabilities for the next five years, this is the conversation you need to hear.
Chapters:
0:00 – Is AI Infrastructure Different?
0:12 – Why It’s Time to Rethink Your IT Architecture
1:20 – The First Difference: Accelerated Compute and GPUs
2:10 – Building Purpose-Built AI Factories
2:42 – The Second Difference: A New Data Architecture
3:09 – Introducing the Knowledge Layer
4:47 – Where Applications Live in an AI Factory
5:37 – Agentic Platforms and the Application Layer
6:09 – The Benefits: Performance, Scale, Security and Observability
7:49 – Bridging AI and Traditional Infrastructure with AI Protocols
8:51 – The Bottom Line: Building for AI First
