AMD’s "Advancing AI 2026" event outlines a strategic pivot toward the "agentic era," where AI evolves from answering questions to executing complex, multi-step tasks. This transition drives a massive surge in compute demand, with inference now consuming 60% of global AI capacity. To meet this, AMD introduces the Helios AI rack, featuring MI455 GPUs and Venice Zen 6 CPUs, designed to scale frontier model performance and agentic workloads. The company also launches ROCm.ai, an agentic platform that leverages AI to automate GPU kernel optimization and software development. Through deep collaborations with industry leaders like OpenAI, Meta, and Cisco, AMD is building a distributed, open-source ecosystem that spans from data centers to personal devices and physical robotics, aiming to capture a $2 trillion accelerator market by 2030 while prioritizing power efficiency and developer productivity.
Sign in to continue reading, translating and more.
Open full episode in Podwise