AMD is expanding its push into enterprise AI, striking a $250 million partnership with Nutanix aimed at building a full-stack AI platform for data centers, clouds and edge sites.
Under the multi-year agreement, AMD will purchase $150 million of Nutanix stock and commit up to $100 million to joint engineering and go-to-market initiatives. The companies plan to develop an integrated AI infrastructure stack optimized for AMD’s EPYC CPUs and Instinct GPUs, with the first jointly developed platform expected in late 2026.
The collaboration positions AMD more aggressively against NVIDIA in the race to supply the hardware underpinning enterprise AI. Today, Nutanix’s AI offerings run on NVIDIA GPUs. By tuning its cloud and Kubernetes platforms to AMD silicon and integrating AMD’s ROCm software ecosystem, Nutanix will add a second accelerator option for customers.
Inference and Agentic AI
The target market is enterprise customers building inference and agentic AI applications. Regulated industries concerned with data sovereignty are a particular focus, as they often prefer to run AI workloads close to where data resides rather than in public clouds.
The deal comes as Nutanix reported solid fiscal second-quarter results. Revenue rose 10 percent year over year to approximately $723 million, surpassing its prior guidance. Annual recurring revenue climbed 16 percent to $2.36 billion. The company added more than 1,000 new customers in the quarter, its highest quarterly total in eight years, bringing its customer count to just under 31,000.
Much of that growth stems from migrations away from VMware following Broadcom’s acquisition of the virtualization giant. Clearly, many enterprises are reassessing their relationships with VMware and exploring alternatives. While some customers are evaluating Red Hat’s Kubernetes-based virtualization tools, Nutanix claims that its migration path is less complex for traditional enterprise workloads.
Yet Nutanix faces challenges. As the quarter progressed, Nutanix executives said they began to see worsening supply chain constraints, particularly around server components. Earlier shortages centered on CPUs. More recently, tight availability of memory and storage has compounded delays.
An Open, Multi-Vendor Stack
Overall enterprise trends seem to support AMD’s move. Across the industry, demand for GPUs remains intense as enterprises expand AI training and inference workloads. Hyperscalers and model developers have reported capacity limits, while chipmakers work aggressively to increase output and develop alternatives to NVIDIA’s leading products.
For AMD, the Nutanix partnership is both a financial investment and a larger foothold in enterprise AI infrastructure. By embedding its accelerators into a widely deployed hybrid cloud platform, AMD gains exposure to customers modernizing data centers and building production-grade AI systems.
For Nutanix, the alliance expands its hardware ecosystem at a moment when AI is becoming crucial to enterprise IT strategy. At this point, AI inference workloads are still in the early stages within corporations. The two companies are gearing up for a world where an open, multi-vendor stack, rather than a vertically integrated approach, will appeal to enterprises seeking flexibility as they scale AI across mixed environments. That appears to be a smart move.
If the partnership delivers as planned, the agreement could be part of a major shift in how enterprise AI infrastructure is built, and which vendors supply the chips that drive it.
