We tour a new NVIDIA B200 cluster by Supermicro for Lambda. This is an AI cluster with thousands of new Blackwell generation GPUs housed at Cologix in Columbus, Ohio. 400GbE and NDR InfiniBand networking, 10s of PB of VAST clustered storage, CPU compute nodes, massive power and cooling, there is a ton to see. This is another STH AI cluster tour. STH Main Site Article: Substack: STH Top 5 Weekly Newsletter: ---------------------------------------------------------------------- Become a STH YT Member and Support Us ---------------------------------------------------------------------- Join STH YouTube membership to support the channel: Professional Users Substack: ---------------------------------------------------------------------- Where to Find The Unit We Purchased Note we may earn a small commission if you use these links to purchase a product through them. ---------------------------------------------------------------------- STH Merch on Spring: ---------------------------------------------------------------------- Where to Find STH ---------------------------------------------------------------------- STH Forums: Follow on Twitter: Follow on LinkedIn: Follow on Facebook: Follow on Instagram: ---------------------------------------------------------------------- Other STH Content Mentioned in this Video ---------------------------------------------------------------------- - Touring a 100,000 GPU AI Cluster: - A fun data center tour: - 2025 AI PCIe GPU Guide: - Inside the NVIDIA H200 AI Server: - Modern AI server power: ---------------------------------------------------------------------- Timestamps ---------------------------------------------------------------------- 0:00 Introduction 1:36 Cooling an AI GPU Cluster 2:13 Powering an AI Cluster 2:51 AI Cluster Facility Design Cold/Hot Aisle 3:20 Storage 4:11 Networking 5:45 Supermicro NVIDIA HGX B200 8 GPU servers 9:59 NVIDIA DGX GB200 NVL72 14:27 Final Words and a Hint at the Next Tour