A Cloud Native Stack From Bare Metal To Tokens for Large-Scale AI Inference
Talkat KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China 2026 · 2026-09-07..09-09·
summary
Platform unifies a mixed fleet of T4, L40S, A100, H200 and B200 GPUs into one managed stack by layering OpenInfra (OIF) at bare metal, CNCF tooling for GPU pooling and multi‑tenant slicing, and PyTorch Foundation serving projects. It runs on ~2,500 GPUs, converts GPU hours into a priceable Token‑as‑a‑Service, cuts GPU inventory by 71% for the same throughput, and delivers OpenAI‑compatible endpoints with provisioning under five minutes.
AI-written summary · about
related talks
12mHow To Run AI At Scale Without Going Broke
30mGreen AI in Cloud Native Ecosystems: Strategies for Sustainability and Efficiency
12mKeynote: AI Infrastructure
33mIBM's Approach to Building a Cloud-Native AI Platform
22mAI @Scale
01h 15mTutorial: Build, Operate, and Use a Multi-Tenant AI Cluster Based Entirely on Open Source
05mOpen Source Infrastructure for the AI Native Era
13mDemocratizing AI with Open Source Compute