The New Stack Podcast CTO Chris Aniszczyk on the CNCF push for AI interoperability
Jan 22, 2026
Chris Aniszczyk, co-founder and CTO of the Cloud Native Computing Foundation, discusses the intersection of AI and cloud-native technology. He explores the similarities between AI agents and microservices, stressing that true 'AI native' systems must be cloud native. Aniszczyk highlights CNCF's initiatives to set AI workload standards, addressing the industry's inconsistencies in GPU and TPU resource management. He also introduces exciting projects like Metal³ for bare-metal Kubernetes and OpenYurt for edge deployments, forecasting a bright future for AI infrastructure.
AI Snips
Chapters
Transcript
Episode notes
AI Native Built On Cloud Native
- AI-native systems should be built on cloud-native foundations like containers, Kubernetes, and observability tools.
- Chris Aniszczyk argues agents resemble microservices but need different scaling and management approaches.
Enforce AI Conformance Across Clouds
- Standardize AI platform capabilities to avoid fragmented, incompatible approaches across clouds and vendors.
- Use CNCF-style conformance to ensure baseline support for accelerators, networking, and storage for AI workloads.
AI Workloads Demand New Platform Primitives
- Kubernetes evolved to support accelerators via features like Dynamic Resource Allocation (DRA) for GPUs and TPUs.
- AI inference also demands specialized networking and storage patterns beyond typical HTTP microservices.