Super Micro Computer Inc. Unveils Next-Generation H15 Server Portfolio

The H15 server portfolio is designed to satisfy the increasing computational demands of various sectors, including cloud computing, enterprise operations, storage solutions, high-performance computing (HPC), and agentic AI workloads. With configurations offering up to 256 cores and 512 threads, these systems are engineered to efficiently handle more concurrent AI agents and accelerate enterprise applications—all while maintaining existing power envelopes.

Vik Malyala, Chief Business Officer at Supermicro, emphasized the significance of these advancements, stating, "The latest AMD powered additions to our DCBBS family deliver the next generation of AI infrastructure, optimized for high performance, rapid scalability, and peak efficiency." He further highlighted Supermicro's commitment to supporting customers in deploying and scaling AI solutions confidently.

The H15 portfolio includes a range of specialized systems tailored to different operational needs:

The flagship dual-socket platform, engineered for enterprise applications and AI inference, features advanced thermal design to support AMD EPYC processors.

This server is designed for cloud-scale environments, built on the Open Compute Project (OCP) Data Center Modular Hardware System (DC-MHS) specification, ensuring compatibility with open data center standards.

A high-density solution featuring a 2U, four-node architecture suitable for scale-out environments, including object storage and virtualization.

This 1OU, two-node compute system maximizes power efficiency for cloud-native and hyperscale deployments using liquid cooling technologies.

High-capacity all-flash storage platforms, optimized for AI data lakes and large-scale analytics, support up to 4.8 PB per system.

The H15 8U SuperBlade platform represents a next-generation architecture designed for HPC, AI inference, and enterprise workloads, available in both air and liquid-cooled configurations.

Supermicro is also broadening its AMD GPU-powered AI infrastructure with the introduction of new PCIe GPU servers and the Supermicro AMD Helios Platform. The two 5U PCIe GPU servers, AS-5126GS-TNRT and AS-5126GS-TNRT2, are optimized for AMD Instinct MI350P GPUs, capable of supporting up to ten GPUs in a standard air-cooled platform. These systems promise exceptional AI acceleration while adhering to existing data center power and cooling parameters.

To enhance connectivity, Supermicro introduces the AMD Pensando Pollara 400 AI NIC, providing high-performance, open Ethernet networking essential for AI infrastructure. This component ensures optimal connectivity for AMD Instinct MI350P-based systems, offering the bandwidth and low latency needed for efficient AI training and inference.

Designed for organizations working with frontier AI models, the Supermicro AMD Helios Platform offers a 72-GPU rack-scale solution that combines liquid-cooled AMD Instinct MI455X GPUs, 6th Gen AMD EPYC processors, and AMD Pensando networking technologies. This platform enables efficient scaling while maximizing performance, energy efficiency, and operational flexibility.

Supermicro’s dedication to providing complete, validated AI infrastructure is evident in its DCBBS offerings, which allow organizations to deploy solutions ranging from individual servers to fully integrated rack-scale systems. The company’s focus on design, manufacturing, and global support services positions it as a pivotal player in accelerating AI adoption, improving energy efficiency, and lowering total cost of ownership.

For those interested in a firsthand look at the new offerings, Supermicro will be showcasing its innovations at the AMD Advancing AI Day 2026, held from July 22-23 at Moscone West in San Francisco.

As the demand for AI infrastructure continues to surge, Super Micro Computer, Inc. is poised to lead the charge with its groundbreaking H15 server portfolio and continued commitment to innovation in technology solutions.