AMD Tech Radar: Building rack-scale AI factories to capture hyperscaler revenue

The single biggest technology story at AMD is its aggressive shift toward building and deploying full-stack, rack-scale AI infrastructure to drive massive revenue growth. To capture this growth, executives confirmed significant investments in both external products and internal efficiency. AMD integrated its acquisition of ZT Systems to build "Helios," a fully integrated AI rack-scale platform combining GPUs, CPUs, and networking to accelerate data center deployments and revenue generation. AMD is also utilizing its chiplet architecture to co-engineer custom MI450-based GPUs specifically tailored for Meta's massive infrastructure build-out.

Internally, AMD is deploying AI to optimize costs and boost efficiency, with executives noting they are "driving AI adoption across all our engineering team" to speed up software coding, hardware verification, and chip design.

Beyond these confirmed initiatives, AMD’s business transformation from a pure chip supplier to a provider of full rack-scale systems—coupled with the divestiture of ZT's manufacturing arm—implies major internal technology upgrades.

This pivot logically requires deploying advanced supply chain and lifecycle management software, alongside complex system-level simulation and testing infrastructure to seamlessly integrate silicon, networking, and liquid cooling at a data center scale.

All sourced directly from AMD's Q4 2025 earnings call / 2025 Investor Day.