Super Micro Computer (trading as Supermicro) has started shipping its H15 server portfolio, targeting data centre operators managing cloud, high-performance computing, and agentic AI workloads.
The hardware family incorporates 6th Gen AMD EPYC 9006 Series processors, formerly codenamed “Venice”, as part of the vendor’s Data Center Building Block Solutions (DCBBS) architecture.
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
By integrating the latest AMD processors, graphics units, and networking into its modular architecture, the vendor reports a 33% increase in processor cores, twice the PCIe bandwidth, and 2.6× higher memory bandwidth compared to earlier AMD generations.
For frontier-model host nodes, these specifications deliver 1.8× higher tokens per second.
The server family encompasses SP7 and SP8 socket systems, covering enterprise GPU units, multi-node platforms, workstations, and edge hardware to supply data centre processing and storage requirements.
Supermicro marketing senior vice president Michael McNerney said: “With the H15 portfolio now shipping, Supermicro and AMD are giving customers a faster, more efficient path to an agentic AI infrastructure.
“Built on our DCBBS architecture and powered by 6th Gen AMD EPYC 9006 processors, these servers deliver substantially higher core density, memory bandwidth, and PCIe throughput.
“Our close engineering collaboration with AMD provides customers the flexibility to scale AI, cloud, and enterprise workloads with confidence.”
The product line spans over 20 configurations, including the dual-socket Hyper platform and the CloudDC server, which conforms to the Open Compute Project Data Center Modular Hardware System specification.
Additional configurations include the liquid-cooled, two-node FlexTwin, the scale-out GrandTwin, the H15 6U SuperBlade in air- and liquid-cooled editions, and 2U Petascale Storage arrays providing up to 4.8 PB of all-flash capacity per unit.
Supermicro noted that the 2U Hyper and CloudDC platforms have achieved multiple SPEC CPU benchmark records.
Supermicro is also collaborating with AMD to deploy rack-scale data centre solutions designated as Helios and Helion.
Helios integrates AMD Instinct MI455X GPUs to accelerate frontier model training and inference throughput, whereas Helion relies on 6th Gen AMD EPYC central processing units for compute flexibility.
Both platforms integrate with the DCBBS framework to support deployments ranging from enterprise agentic systems to hyperscale facilities.
Last month, Supermicro expanded its portfolio of Open Compute Project (OCP) accepted systems, racks, and cooling products targeted at AI data centres.
