AMD and Microsoft Just Went All-In on Azure’s AI Future

The AI infrastructure race just got another major move. AMD and Microsoft have expanded their long-term partnership, and this time it covers the full stack — GPUs, CPUs, networking, and software — all built to power Azure’s next generation of AI workloads.

Azure’s Big Announcement

At the center of this expansion is the AMD Helios Rackscale Solution, which Microsoft will deploy to run frontier AI model inference for itself, its AI customers, and Azure AI services. Helios brings together AMD Instinct MI455X GPUs, EPYC “Venice” CPUs, Pensando networking, and ROCm software into one open, integrated platform built for large-scale AI training and inference. Shipments to customers, including Microsoft, begin in the second half of 2026.

AMD CEO Lisa Su called this an extension of years of infrastructure work between the two companies, while Microsoft CEO Satya Nadella framed it as giving customers the performance and flexibility needed to build the next wave of AI applications.

New Azure VM Series

Azure is introducing two new virtual machine series, both powered by 6th Gen AMD EPYC “Venice” processors:

• Azure HDv2 — built for agentic AI and data pipeline workloads
• Azure HXv2 — designed for semiconductor design workflows

Together, these expand Azure’s AMD EPYC lineup across AI, data, and engineering use cases.

Networking Gets an Upgrade Too

The partnership isn’t limited to compute. Microsoft is broadening its use of AMD Pensando DPUs and integrating them with Azure Boost, aiming to improve networking performance, efficiency, and connection processing across Azure’s infrastructure at scale.

Deal Highlights

Why This Matters

As AI demand keeps climbing, cloud providers are racing to diversify beyond a single chip supplier. This deal gives Azure customers more choice for training and inference workloads, while reinforcing AMD’s push into large-scale AI infrastructure — a space it has been steadily expanding, as outlined in AMD’s company history on Wikipedia.

For more updates on AI infrastructure, chipmakers, and cloud computing, check out TechnoSports for regular coverage.

FAQs