Supermicro is a global leader in application-optimized total IT solutions, delivering servers, storage, networking and rack-scale systems to enterprises, cloud providers and AI data centers worldwide. Its building-block approach lets organizations configure platforms precisely for their workloads, from the edge to large-scale AI infrastructure. Intel and Supermicro are collaborating to deliver optimized, validated systems built on Intel® Xeon® 6 processors to deploy and scale AI inference, confidential computing and agentic AI.
As AI moves into regulated and multi-tenant environments, the infrastructure has to protect models, prompts and data while they are in use. Supermicro's reference architecture for confidential AI pairs Intel® Trust Domain Extensions (Intel® TDX) on Xeon 6 hosts with NVIDIA HGX™ B300 GPUs in Confidential Computing mode. Trust domains isolate the workload from the host, the hypervisor and other tenants, and an encrypted bounce buffer protects data on its way between CPU and GPU. The design is validated end to end on Confidential Containers and Red Hat® OpenShift® Sandboxed Containers, with CPU and multi-GPU attestation verified through Trustee, NVIDIA NRAS and Intel® DCAP, from one to eight GPUs per confidential pod.
Read the Supermicro white paper on Intel TDX, or visit our partner page to learn more about how Intel and Supermicro work together.