Source
Building Infrastructure for an AI World in Motion
AMD argues that AI infrastructure must be designed for adaptability, economy, and openness to support rapidly evolving workloads like agentic AI.
Why it matters
Companies risk investing in rigid IT foundations that cannot adapt to changing AI requirements over the next five to seven years.
The big picture
This provides the infrastructure perspective for agentic AI, a technology previously mentioned in AMD's collaboration with F-Secure.
The details
- Agentic AI steps include retrieving inputs, calling tools, executing code, and coordinating sub-agents. - AMD's hardware range spans EPYC CPUs, Instinct GPUs, and the Helios rackscale solution. - Cost for agentic AI should be measured per finished task rather than per token.
Show entities and relationshipsHide entities and relationships
In this article
Key connections
AMD owns AMD Radeon
AMD has invested across the entire stack... local AI on AMD Radeon™
AMD Instinct uses fine-tuning
Large language model inference and fine-tuning on AMD Instinct™ PCIe cards
Get the weekly recap
The stories like this one, picked and explained — once a week, straight to your inbox.