The NVIDIA GB200 NVL72 is a liquid-cooled, data-centre-class AI server delivering unprecedented performance for generative AI, large language models and HPC workloads. With a 72-GPU NVLink domain, the system acts as a unified accelerator delivering up to 1,440 PFLOPS FP4, 720 PFLOPS FP8, and supports massive 13.4 TB HBM3e memory and 130 TB/s high-speed interconnect. Designed for real-time trillion-parameter inference (30× H100) and large-scale training (4× H100), this rack-scale solution enables hyperscale AI infrastructure with extreme throughput and efficiency.

The NVIDIA GB200 NVL72 is a liquid-cooled, data-centre-class AI server delivering unprecedented performance for generative AI, large language models and HPC workloads. With a 72-GPU NVLink domain, the system acts as a unified accelerator delivering up to 1,440 PFLOPS FP4, 720 PFLOPS FP8, and supports massive 13.4 TB HBM3e memory and 130 TB/s high-speed interconnect. Designed for real-time trillion-parameter inference (30× H100) and large-scale training (4× H100), this rack-scale solution enables hyperscale AI infrastructure with extreme throughput and efficiency.

The HPE NVIDIA GB300 NVL72 AI Server is a fully liquid-cooled, rack-scale solution combining 36 NVIDIA Grace CPUs and 72 NVIDIA Blackwell Ultra GPUs into a single high-performance NVLink domain. Designed for next-generation AI reasoning, large language models, and data-intensive workloads, it delivers unmatched density, bandwidth, and throughput for hyperscale AI deployments.

The HPE NVIDIA GB200 NVL72 is a high-performance rack-scale AI platform designed for generative AI, large language models, and HPC workloads. It combines 72 NVIDIA Blackwell GPUs and 36 NVIDIA Grace CPUs with a massive NVLink GPU domain and advanced liquid cooling to deliver unparalleled training and real-time inference capabilities.

High-performance AI compute platform featuring 8× NVIDIA Blackwell Ultra B300 GPUs, delivering exceptional acceleration for AI training, inference, and HPC workloads in a scalable air-cooled design.

The NVIDIA GB200 NVL72 is a liquid-cooled, data-centre-class AI server delivering unprecedented performance for generative AI, large language models and HPC workloads. With a 72-GPU NVLink domain, the system acts as a unified accelerator delivering up to 1,440 PFLOPS FP4, 720 PFLOPS FP8, and supports massive 13.4 TB HBM3e memory and 130 TB/s high-speed interconnect. Designed for real-time trillion-parameter inference (30× H100) and large-scale training (4× H100), this rack-scale solution enables hyperscale AI infrastructure with extreme throughput and efficiency.

The NVIDIA GB300 NVL72 is engineered as a turnkey AI infrastructure solution, combining cutting‑edge hardware and networking to power next‑generation reasoning workloads. With 72 Blackwell Ultra GPUs, 36 Grace CPUs, up to ~37 TB of fast memory, and a 130 TB/s NVLink fabric, it enables enterprises to deploy AI factories at hyperscale. Whether for large language model inference, real‑time video generation or agentic systems, the GB300 NVL72 delivers up to 50× the output of previous generation platforms. Fully liquid‑cooled for efficiency and bundled with NVIDIA Mission Control and a high‑performance networking stack (Quantum‑X800 InfiniBand / Spectrum‑X Ethernet) via ConnectX‑8 SuperNICs.

The Lenovo ThinkSystem SR655 V3 is a 1-socket 2U rack server powered by the 5th Gen AMD EPYC 9005 processor, delivering up to 160 cores and 320 threads. It supports DDR5 memory up to 6400 MHz, offers expansive GPU capacity, and features the latest PCIe Gen5 for high-speed I/O. Ideal for AI inference, VDI, analytics, and software-defined storage workloads.​

​The Lenovo ThinkSystem SR850 V3 Controlled GPU Server delivers enterprise-class performance and reliability for demanding AI, HPC, and business-critical applications. With support for up to four 4th Gen Intel® Xeon® Scalable processors, up to 64 DIMM slots for 16TB DDR5 memory, flexible storage up to 1.4PB, and GPU acceleration, this 2U rack server provides exceptional scalability and efficiency. Ideal for SAP HANA, AI training, big data analytics, and high-transaction workloads.​

​The Lenovo ThinkSystem SR665 V3 Controlled GPU Server is a 2U rack server engineered for AI, HPC, and data-intensive workloads. Featuring dual AMD EPYC 4th/5th Gen processors with up to 160 cores, support for 24 TruDDR5 DIMM slots, high-capacity storage, advanced GPU support, and enterprise-grade security, it provides exceptional scalability, performance, and reliability for modern data centers. ​

​Lenovo ThinkSystem SR675 V3 Rack Server combines AMD EPYC™ processors and NVIDIA GPU acceleration in a versatile 3U form factor. Designed for AI, deep learning, and HPC workloads, it offers flexible storage, high-speed memory, and advanced cooling options to maximize performance and efficiency.

The Supermicro SYS-422GA-NBRT-LCC GPU SuperServer delivers extreme performance for mission-critical AI and HPC. Featuring dual Intel® Xeon® Scalable processors, NVIDIA® HGX™ B200 with 8 GPUs, DDR5 memory, PCIe Gen5, and advanced liquid cooling, this 4U rack system is engineered for training trillion-parameter AI models, data analytics, and next-gen scientific computing with maximum efficiency and scalability.

NVIDIA HGX A100 8-GPU and 4-GPU accelerators powered by NVIDIA A100 Tensor Core GPUs with NVLink; AMD Instinct MI100 accelerators; broad choice of PCIe GPU for HPC or AI.

HPE ProLiant Compute XD685 systems powered by NVIDIA are optimized for natural language processing, large language model training and multimodal training.​

HPE ProLiant Compute XD685 systems powered by NVIDIA are optimized for natural language processing, large language model training and multimodal training.​

The HPE ProLiant Compute DL380a Gen12 server is a rack-optimized, 4U 2P solution that delivers breakthrough performance with advanced GPU accelerators on an ultra-scalable architecture.​

As part of NVIDIA AI Computing by HPE, HPE ProLiant Compute DL384 Gen12 with NVIDIA GH200 NVL2 handles next-level performance and scale-out fine tuning and inference with RAG.​

High-density 5U server with dual Intel Xeon CPUs, up to 8x GPUs, and HPE's Silicon Root of Trust security for AI and HPC workloads.​

HPE Cray XD670 CTO Server – A 5U rack-mounted powerhouse featuring dual Intel® Xeon® processors, up to eight NVIDIA® Tensor Core GPUs, and optional direct liquid cooling, designed for scalable AI training and deep learning performance.

High-density 5U server with dual Intel Xeon CPUs, up to 8x GPUs, and HPE's Silicon Root of Trust security for AI and HPC workloads.​