NVIDIA AI Podcast Building AI Factories: How Red Hat and NVIDIA Turn Enterprise Data Into Intelligence - Ep. 293
57 snips
Mar 12, 2026 Chris Wright, Red Hat CTO working on enterprise AI platform and governance, and Justin Boitano, NVIDIA VP overseeing AI infrastructure and blueprints. They explore the five-layer AI factory stack, agentic systems that plan and act, production-grade governance and security, hybrid cloud inference strategies, and starting small with validated designs and iterative evals.
AI Snips
Chapters
Transcript
Episode notes
Five Layer Stack Powers AI Factories
- AI factories are a five-layer stack from data-center hardware up through orchestration, models, agents, and apps that turn data into business intelligence.
- Justin Boitano describes the base as rack-scale chips and power, then orchestration, models, and agent/apps to produce use-case outcomes like faster product development.
Builder Tools Need Enterprise Guardrails
- Enterprises must pair cutting-edge models and builders' enthusiasm with enterprise-grade controls like access controls, audit trails, and security.
- Chris Wright emphasizes that builder-focused tools (e.g., OpenClaw) lack enterprise guardrails, so Red Hat and NVIDIA integrate security into the stack.
Agents Move From Code To Knowledge Work
- Agentic systems are moving from coding-first use cases into broad enterprise knowledge work with hybrid model architectures for cost and privacy gains.
- Justin Boitano cites enterprise search and hybrid private/open models delivering up to 30x cost reduction in blueprints.