Category: GPU

A versatile 5U air-cooled GPU server featuring 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 768GB total GDDR7 memory, dual Intel Xeon 6952P processors (192 cores), 1.5TB DDR5-6400 ECC RAM, and dual-root PCIe 5.0 architecture. Purpose-built for AI inference, LLM fine-tuning, generative AI, 3D rendering, VDI, and HPC workloads requiring dense PCIe GPU compute with enterprise-grade reliability.

The HPE Cray XD670 is a 5U air-cooled GPU server featuring 8 NVIDIA H100 SXM5 80GB GPUs with 900 GB/s NVLink interconnect, dual Intel Xeon Platinum 8462Y+ processors, 2TB DDR5 ECC memory, and 3.2 Tb/s of InfiniBand NDR fabric bandwidth via 8 ConnectX-7 adapters. Purpose-built for large-scale AI training, LLM fine-tuning, generative AI, and HPC workloads requiring maximum GPU density and interconnect performance.

The Lenovo ThinkSystem SR780a V3 is a liquid-cooled 5U GPU server featuring 8 NVIDIA HGX B200 SXM6 GPUs with 1.44TB of HBM3e memory, dual Intel Xeon Platinum 8570 processors, 2TB DDR5-5600 ECC RAM, and 8x 400 Gb/s NDR InfiniBand networking via ConnectX-7. Purpose-built for large-scale AI training, LLM inference, generative AI, and HPC workloads requiring maximum compute density, GPU interconnect bandwidth, and energy-efficient liquid cooling.

A high-density air-cooled 8U AI server featuring 8 NVIDIA HGX B300 Blackwell Ultra GPUs with 2.3TB of HBM3e memory, dual Intel Xeon 6767P processors, 3TB DDR5 ECC memory, and 800 Gb/s OSFP networking. Purpose-built for large-scale AI training, LLM reasoning, generative AI, and HPC workloads requiring maximum compute density and interconnect bandwidth.

A quick-ship 10U AI server featuring 8 NVIDIA HGX B200 Blackwell SXM GPUs with 1.44TB of HBM3e memory, dual Intel Xeon 6960P processors (144 cores), 2TB DDR5-6400 ECC memory, and 400 Gb/s OSFP networking via ConnectX-7. Purpose-built for large-scale AI training, LLM reasoning, generative AI, and HPC workloads requiring maximum compute density and GPU interconnect bandwidth.

A high-density air-cooled 8U AI server featuring 8 NVIDIA HGX B300 Blackwell Ultra SXM GPUs with 2.3TB HBM3e, dual Intel Xeon 6767P processors, 3TB DDR5 ECC memory, and 8x 800 Gb/s ConnectX-8 SuperNIC networking. Purpose-built for large-scale AI training, LLM reasoning, generative AI, and HPC workloads requiring maximum compute density, interconnect bandwidth, and enterprise reliability.

The Supermicro SYS-821GE-TNHR is an air-cooled 8U GPU server featuring 8 NVIDIA H100 80GB SXM5 GPUs on the HGX platform with 900 GB/s NVLink per GPU, dual Intel Xeon 8570 processors (56 cores each), 2TB DDR5-5600 ECC memory, and 8x NDR 400 Gb/s OSFP networking via ConnectX-7. Purpose-built for large-scale AI training, LLM inference, generative AI, and HPC workloads requiring maximum GPU interconnect bandwidth and dense compute.

The Supermicro AS-8126GS-TNMR is an air-cooled 8U GPU server featuring 8 AMD Instinct MI350X OAM accelerators with 2.3TB total HBM3e memory, dual AMD EPYC 9575F processors (128 cores total), 2.25TB DDR5-6400 ECC memory, and 8x 400 Gb/s NDR InfiniBand/Ethernet OSFP networking via NVIDIA ConnectX-7. Purpose-built for large-scale AI training, LLM inference, generative AI, and HPC workloads requiring massive GPU memory capacity and high-bandwidth interconnect.

A high-density liquid-cooled 8U AI server featuring 8 NVIDIA HGX B300 GPUs, dual Intel Xeon processors, DDR5 memory, and 800 GbE networking. Designed for large-scale AI training, LLMs, and HPC workloads requiring maximum performance and bandwidth.