Powered by AMD EPYC™ 9006 Series Processors
The 6th Gen AMD EPYC™ Server CPUs deliver continued leadership for the modern data center. Engineered to handle every stage of the agentic workflow, this new architecture provides the exceptional compute density and efficiency required to scale more concurrent agents, maximize AI host-node throughput, and drastically expand memory bandwidth.
Built on the cutting-edge 2nm process node and Zen 6 architecture, the AMD EPYC™ 9006 Series Server CPUs introduce next-generation I/O capabilities to eliminate data center bottlenecks. Featuring PCIe® Gen 6 for double the bandwidth per link and CXL 3.1 for cost-effective memory expansion, this architecture establishes an entirely new baseline for data throughput and accelerator orchestrations.
The AMD EPYC™ 9006 Series Server CPUs scale up to 256 cores to deliver massive thread density, allowing organizations to maximize agentic capacity per watt and per dollar. By providing up to 1.6TB/s of memory bandwidth and optimized CPU-to-GPU connectivity, this broad, workload-optimized portfolio drives absolute performance for demanding enterprise applications, backend databases, and real-time model inference.
Purpose-built for cloud infrastructure, HPC, and AI host nodes, the AMD EPYC™ 9006 SP7 Server CPU maximizes throughput and performance per watt. With up to 256 cores and 512 threads, it enables massive virtual machine consolidation and concurrent agent sandbox execution. High memory bandwidth ensures AI accelerators remain fully saturated, helping to lower total cost of ownership for large-scale deployments by fitting more capacity into existing power envelopes.
Designed for enterprise environments, the AMD EPYC™ 9006 SP8 Server CPU balances compute, memory, and I/O to optimize performance per system dollar. Flexible options from 8 to 128 cores allow organizations to precisely match capacity to general IT, virtualization, edge, and telco deployments. By improving power efficiency and lowering platform costs, the SP8 provides an adaptable, cost-effective foundation for the backend databases and services that support agentic workflows.
