AMD announced on November 10, 2026, the general availability of its 6th Gen AMD EPYC server SP7 processors, built on the new Zen 6 architecture. Systems powered by the new CPUs are now shipping from leading server OEMs and ODMs, and cloud instances from Microsoft Azure and Oracle Cloud Infrastructure will be generally available soon, the company said in its Santa Clara announcement.
The launch is AMD's clearest statement yet that the CPU is becoming more important, not less, as AI moves into agents. A single agentic workflow can run agents, call models, retrieve data, execute code and interact with the databases and applications enterprises already use. AMD is positioning the new portfolio as workload-optimized for exactly that mix, spanning agentic AI, enterprise services, cloud computing, high-performance computing and general-purpose computing, while preserving a common x86 software, security and operational foundation.
"AI is significantly expanding the role of the CPU across every layer of the data center," said Dan McNamara, senior vice president and general manager for Compute and Enterprise AI at AMD. "With 6th Gen AMD EPYC, customers can standardize on the proven x86 foundation while choosing a workload-optimized solution to run more agents, keep accelerated infrastructure fed, and power the business-critical applications they depend on."
At the top of the stack, AMD called the EPYC 9996 the world's best general-purpose processor. The company said a system based on that chip delivers up to 3.7 times more performance than an Nvidia Vera CPU-based system and up to 2.2 times the performance and 2.0 times the performance per dollar of an Intel Xeon 6980P system, citing SPECrate 2017 integer results and its own estimates for the competing systems. AMD presented that as more applications, transactions and users per server, within existing power and footprint limits.
For agentic AI specifically, AMD described three distinct CPU roles: general-purpose services, GPU host nodes, and agent sandboxes. Its high-frequency host-node processors reach up to 5.3 GHz and, AMD said, can deliver up to 1.8 times the large language model output throughput of the previous EPYC generation. Its high-density processors, the company added, can provide up to about 3.3 times the agents-per-CPU-watt advantage over Nvidia Vera and about 2.5 times over Intel Xeon 6980P for hosting agent sandboxes.
The company also extended the story from servers to racks with its Helion agentic AI racks. Helion Liquid is designed to sit alongside the Helios rack-scale solution in large deployments, Helion Air is for expanding agent capacity in existing air-cooled data centers, and Helion Air Inference combines CPU execution and local GPU inference in one rack. Some Helion configurations reach up to 73,728 threads in a single rack, AMD said.
Ecosystem support was central to the announcement. Meta is working with AMD as a lead customer and has been validating 6th Gen EPYC platforms ahead of deployment at scale. Microsoft Azure said HDv2 is now in preview for AI data processing, agentic AI and HPC, with HXv2 for electronic design automation and technical computing, and the Azure ND MI455X v7, based on AMD Helios, for production-scale AI inference. Oracle's OCI Compute E7 instances will also be powered by the new processors.
AMD noted that its 5th Gen EPYC line became the fastest-ramping and highest-volume EPYC generation in its history. Whether the 6th generation repeats that will depend on how quickly the shipping systems and the forthcoming Azure and OCI instances convert into production agent workloads, but the direction is clear: as agents multiply, the host CPU, the sandbox and the services around the accelerator all need to scale together.