Uploaded August 2026 | Updated September 2026, 2 weeks ago

---

See a full end-to-end Agentic RAG pipeline running on Supermicro hardware with AMD Instinct MI350X GPUs, built on AMD’s official Enterprise AI Solution Blueprints.

In this video we walk through:

▸Traditional RAG vs Agentic RAG

▸Why memory capacity and bandwidth matter for agentic workloads
The Supermicro AS-8126GS-TNMR 8U system with AMD Instinct MI350X GPUs

▸AMD’s open Solution Blueprints and AIMs

▸Architecture overview: LangGraph agent + Model Context Protocol (MCP)

▸Key components: MCP Server, ChromaDB, Gradio UI, embedding service, and vLLM

This is a practical demonstration of a modular, observable agentic RAG system with document ingestion, iterative retrieval, relevance grading, and live hardware metrics via Grafana + Prometheus.

Resources & Links:

▸AMD Enterprise AI Solutions Blueprints: enterprise-ai.docs.amd.com/en/latest/solution-blueprints/overview.html

▸AMD Solution Blueprint for Agentic RAG GitHub: github.com/amd-enterprise-ai/solution-blueprints/tree/main/solution-blueprints/agentic-rag

▸Supermicro MI350X: supermicro.com/en/products/system/gpu/8u/as-8126gs-tnmr

▸Supermicro Jumpstart for MI350X: learn-more.supermicro.com/mi350x

Timestamps:
0:00 – Hook & Intro
0:35 – Traditional RAG
1:05 – Agentic RAG
1:45 – AI Agents & Enterprise Value
2:15 – Hardware (Supermicro + MI350X)
2:55 – AMD Solution Blueprints
3:20 – Architecture Overview
5:45 – Key Takeaways
6:20 – Resources & CTA

#Supermicro #AMD #AI #EnterpriseAI