As artificial intelligence workloads expand rapidly, no single infrastructure strategy is sufficient to manage them all. With the increasing volume of models, new agent-driven tasks, and a rising demand for computational resources, there is a pressing need for specialized approaches throughout the technology stack. To address this challenge, Microsoft is continuously upgrading Azure’s infrastructure, notably by integrating AMD’s cutting-edge AI and high-performance computing (HPC) solutions.
Microsoft’s strategy for AI infrastructure is crafted to accommodate the diverse methods by which AI systems are developed and operated. By collaborating closely with industry leaders like AMD, along with leveraging its own custom-built silicon and systems, Microsoft offers its clients a robust, open, and varied platform. This platform aims to deliver optimal results in terms of performance, cost-effectiveness, and energy efficiency.
As part of its ongoing partnership with AMD, Microsoft is adopting AMD’s latest Helios AI platform and the next-generation EPYC datacenter processors into Azure. These advanced technologies are set to power three new Azure offerings: HDv2 virtual machines (VMs) for data processing, HXv2 VMs for electronic design automation (EDA), and ND MI455X v7 VMs for AI inference workloads.
Expanding capabilities for AI, data systems, and chip design
With these enhancements, Microsoft aims to expand its infrastructure capabilities in the areas of AI inference, AI data systems, and chip design.