Supermicro Unveils Next-Generation H15 Servers Powered by 6th Gen AMD EPYC CPUs

Super Micro Computer, Inc. has introduced its latest H15 server portfolio, engineered to meet the evolving demands of AI, enterprise, storage, and 5G/Edge IT environments. These new servers are powered by the 6th Gen AMD EPYC 9006 Series processors and are optimized for next-generation GPUs, including AMD Instinct, with advanced connectivity provided by AMD Pensando networking solutions.

Breakthrough Performance for Modern Workloads

The H15 systems deliver up to 256 cores and 512 threads, offering a significant 1.7x generational CPU performance improvement. With expanded memory and I/O bandwidth, these servers provide industry-leading compute density, enabling organizations to run more concurrent AI agents, accelerate enterprise applications, and maximize host-node performance—all while maintaining efficient power consumption.

These advancements position the H15 portfolio as a robust solution for cloud computing, enterprise workloads, high-performance computing (HPC), storage, and agentic AI applications.

Comprehensive Server Portfolio for Diverse IT Needs

The H15 server lineup includes purpose-built systems tailored for a wide range of deployment scenarios:

  • Hyper: Dual-socket platform designed for enterprise applications, AI inference, virtualization, and cloud workloads, featuring advanced thermal management for high-performance AMD EPYC processors.
  • CloudDC: Single or dual-socket servers built for cloud-scale environments, compliant with Open Compute Project (OCP) Data Center Modular Hardware System (DC-MHS) standards for open data center compatibility.
  • GrandTwin: High-density 2U, four-node architecture optimized for scale-out environments such as object storage, virtualization, cloud services, and HPC.
  • FlexTwin: 1OU, two-node, high-density dual CPU system designed for cloud-native and hyperscale deployments, featuring liquid cooling for maximum efficiency.
  • Petascale Storage: 1U and 2U all-flash storage platforms supporting up to 4.8 PB per system, ideal for software-defined storage, AI data lakes, large-scale analytics, and HPC.
  • SuperBlade: The H15 8U 10 SuperBlade offers next-generation rack-scale architecture for HPC, AI inference, agentic AI, and enterprise compute workloads, supporting both air and liquid-cooled configurations for maximum density and performance.

Expanding AMD GPU-Powered AI Infrastructure

Supermicro is further enhancing its AMD GPU-powered AI infrastructure with new PCIe GPU servers and the rack-scale Supermicro AMD Helios Platform. These solutions, showcased at Computex 2026, provide flexible deployment options for enterprise AI inference and large-scale AI training.

5U PCIe GPU Servers with AMD Instinct MI350P

The Supermicro AS-5126GS-TNRT and AS-5126GS-TNRT2 servers are designed to maximize the performance of AMD Instinct MI350P PCIe GPUs. Supporting up to ten GPUs in a standard 5U air-cooled chassis, these systems deliver powerful AI acceleration while fitting seamlessly into existing data center power and cooling infrastructures. With up to 144 GB of HBM3e memory per GPU and support for low-precision AI formats, organizations can accelerate both AI inference and training, improve infrastructure efficiency, and reduce total cost of ownership.

High-Performance Networking with AMD Pensando Pollara 400 AI NIC

The AMD Pensando Pollara 400 AI NIC delivers high-bandwidth, low-latency Ethernet networking for AI infrastructure. This open networking solution enables efficient front-end, storage, and scale-out connectivity for AMD Instinct MI350P-based systems, supporting scalable AI clusters from single servers to multi-rack deployments using standard Ethernet infrastructure.

Supermicro AMD Helios Platform: Rack-Scale AI for Frontier Models

For organizations deploying advanced AI models, Supermicro and AMD have collaborated on the Supermicro AMD Helios Platform—a 72-GPU, rack-scale solution designed for large-scale AI training and high-throughput inference. This liquid-cooled platform integrates AMD Instinct MI455X GPUs, 6th Gen AMD EPYC processors, AMD Pensando networking, and the AMD ROCm software stack to deliver open, high-performance AI infrastructure. The platform supports deployments of all sizes, enabling efficient scaling, maximum performance, and operational flexibility.

Integrated, Validated AI Infrastructure

Supermicro’s Data Center Building Block Solutions (DCBBS) bring together advanced server, storage, networking, and cooling technologies into complete, validated AI infrastructure. This approach enables organizations to deploy solutions ranging from individual servers to fully integrated rack-scale and data center-level systems. With a focus on design excellence, manufacturing quality, liquid cooling, networking, software, and global support, Supermicro empowers customers to accelerate AI adoption, reduce deployment time, improve energy efficiency, and lower total cost of ownership.