Description

In this episode of The Hitchhiker’s Guide to IT, host Michelle Dawn Mooney sits down with Justyn Hornor, CEO of I/ONX, to explore how heterogeneous compute is reshaping enterprise infrastructure. With two decades of experience across AI, ML systems, and hardware optimization, Justyn shares why a flexible, composable approach to infrastructure is essential for scaling high-performance and AI workloads.

The conversation covers:

• Why traditional CPU-GPU models are no longer enough for AI-driven enterprises
• How heterogeneous compute unlocks performance, power efficiency, and cost savings
• A real-world use case where pre-processing on ASICs cut GPU needs by 40%
• The orchestration layer needed to manage diverse chips across racks and data centers
• What IT leaders can do to start small without ripping and replacing
• How composable infrastructure future-proofs organizations for evolving workloads

Whether you're training complex models, optimizing inference, or just trying to get more from existing infrastructure, this episode delivers practical insights for navigating today’s compute demands.

Listen in to learn how to move beyond vendor lock-in, reduce energy consumption, and build an infrastructure that grows with your AI strategy.