Operating Frontier Intelligence at Scale
Talkat KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China 2026 · 2026-09-07..09-09·
summary
Machine workloads shift from training to inference with commits jumping from ~1.4B to 2.9B in months and web traffic driven more by bots and agents, causing token-heavy, always-on inference demand. Kubernetes already runs generative AI for ~66% of users; CNCF lists 30+ certified AI platforms and reports ~1.75M cloud native contributors in China. Projects like Karmada (graduated) and Hammy (incubating) target multi-GPU and ASIC orchestration to improve GPU utilization, backed by Nvidia’s $4M GPU contribution.
AI-written summary · about
chapters (8)
- 0:00From model to scale
- 1:02Inference workloads
- 3:54Agentic workloads
- 5:56CNCF growth
- 7:20Carmada and Hammy
- 13:34AI platforms on Kubernetes
- 16:43Community contribution
- 19:11Cross-organization collaboration
related talks
15mKeynote: From Platforms to AI Factories. Has Kubernetes Solved It?
06mAI Inference at the Edge: Beyond the Hype
09mKeynote: Kubernetes for GenAI: The New Operating System of the AI-First World
10mKeynote: Enabling AI with Cloud Native
12mKeynote: AI Infrastructure
27mHarnessing Kubernetes for Scalable AI/ML Workloads
28mScaling AI Infrastructure
30mGreen AI in Cloud Native Ecosystems: Strategies for Sustainability and Efficiency