- NVIDIA RTX PRO 2000 Blackwell 16GB NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / CAD
A half-height, half-length professional GPU with 16GB GDDR7 at just 70W and no auxiliary power connector — it fits compact workstations and rack chassis alike. Built for CAD, product design, content c
- NVIDIA RTX PRO 4000 Blackwell 24GB NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / RENDERING
A single-slot professional GPU with 24GB GDDR7 at 140W. The best density option when you need several cards in one chassis or space is tight.
- NVIDIA RTX PRO 4500 Blackwell 32GB NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / RENDERING
32GB GDDR7 at 200W with 896 GB/s of memory bandwidth. The workhorse for large 3D scenes, simulation and running mid-sized AI models locally.
- NVIDIA RTX PRO 5000 Blackwell 48GB NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / AI INFERENCE
48GB GDDR7 at 300W, and it can be partitioned into two 24GB instances for separate workloads. A high-end choice for design, data science and simulation.
- NVIDIA RTX PRO 5000 72GB Blackwell NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / AI INFERENCE
Same GPU as the 48GB model with memory increased to 72GB GDDR7, still at 300W. Pick this one for larger local models, or to split a single card into two 36GB instances.
- NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 96GB NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / AI INFERENCE
96GB GDDR7 held to a 300W envelope in a standard dual-slot size. When you need several 96GB cards in one workstation, this is the version the power and thermals allow.
- NVIDIA RTX PRO 6000 Blackwell Workstation Edition 96GB NVIDIA / Leadtek · PROFESSIONAL GPU / WORKSTATION / AI INFERENCE
The fastest single workstation GPU available today: 96GB GDDR7, 600W, double-flow-through cooling. For local fine-tuning of large models, film-grade rendering and the heaviest simulation work.
- NVIDIA RTX PRO 6000 Blackwell Server Edition 96GB NVIDIA / Leadtek · DATA CENTER GPU / AI INFERENCE / LLM
The data center edition: passively cooled with a configurable 400–600W power envelope, designed to be deployed eight per server. Handles generative AI inference, VDI and video workloads, with confiden
- NVIDIA H200 NVL 141GB HBM3e NVIDIA · DATA CENTER GPU / AI TRAINING / LLM
A data center accelerator with 141GB of HBM3e and 4.8 TB/s of bandwidth in a dual-slot, air-cooled PCIe form factor. Two- or four-way NVLink bridging makes it the mainstay for large language model tra
- NVIDIA HGX™ B300 NVL8 Liquid-Cooled AI Server (4U16X-GNR2/DLC B300) ASRock Rack · 4U · AI TRAINING / LLM / INFERENCE / HPC / LIQUID COOLING / AI FACTORY
The 4U16X-GNR2/DLC B300 is an ASRock Rack 4U direct-liquid-cooled AI server with the NVIDIA HGX B300 NVL8 GPU module on a dual-socket Intel Xeon 6 (Granite Rapids) platform. It shares the same liquid-
- NVIDIA HGX B300 NVL8 AI Server ASRock Rack · 8U · TRAINING / INFERENCE / LLM / HPC / AI FACTORY
The 8U16X-GNR2 B300 is an ASRock Rack 8U rack-mount AI training server built around the NVIDIA HGX B300 NVL8 GPU module with NVIDIA NVLink interconnect, delivering a complete eight-GPU Blackwell Ultra
- NVIDIA HGX B300 NVL8 Liquid-Cooled AI Server (4U16X-TURIN2/DLC B300) ASRock Rack · 4U · AI TRAINING / LLM / INFERENCE / HPC / LIQUID COOLING / AI FACTORY
The 4U16X-TURIN2/DLC B300 is an ASRock Rack 4U direct-liquid-cooled AI server with the NVIDIA HGX B300 NVL8 GPU module on a dual-socket AMD EPYC Turin (9005 series) platform. Liquid cooling compresses
- NVIDIA HGX B300 NVL8 AI Server (AMD EPYC Turin Platform) ASRock Rack · 8U · TRAINING / INFERENCE / LLM / HPC / AI FACTORY
The 8U16X-TURIN2 B300 is an ASRock Rack 8U rack-mount AI training server on a dual-socket AMD EPYC Turin (9005 series) platform, paired with the NVIDIA HGX B300 NVL8 GPU module and NVIDIA NVLink inter
- NVIDIA HGX B200 8-GPU AI Server (Synthetic Mode DLC) 4U · AI TRAINING / LLM / INFERENCE / HPC / LIQUID COOLING
The 4U8X-GNR2/DLC SYN B200 is a 4U rack-mount liquid-cooled AI server carrying the NVIDIA HGX B200 8-GPU module with NVIDIA NVSwitch for full-speed GPU-to-GPU interconnect. Direct liquid cooling (DLC)
- NVIDIA HGX H200 8-GPU AI Server (ZutaCore HyperCool Edition) 6U · AI TRAINING / LLM / INFERENCE / HPC / LIQUID COOLING
The 6U8X-EGS2/ZC SYN H200 is a 6U AI server using ZutaCore HyperCool two-phase liquid cooling with the NVIDIA HGX H200 8-GPU module and NVIDIA NVSwitch. ZutaCore uses waterless two-phase direct-contac
- NVIDIA HGX H200 8-GPU AI Server (Synthetic Mode) ASRock Rack · 6U · AI TRAINING / LLM / INFERENCE / HPC / GENERATIVE AI
The 6U8X-EGS2 SYN H200 is an ASRock Rack 6U air-cooled AI server with the NVIDIA HGX H200 8-GPU module and NVIDIA NVSwitch. The H200 offers 141GB of HBM3e memory and 4.8 TB/s of bandwidth, making it e
- MiTAC G4520G6U2BC-N AI GPU Server MiTAC · 4U · AI TRAINING / LLM / INFERENCE / GPU COMPUTING / HPC
The MiTAC G4520G6U2BC-N is a 4U rack-mount GPU server that accepts up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition professional GPUs at 600W each, using a PCIe Gen 5 switch to provide eight f
- B8261T85E8HR-2T-N-HE (TN85-B8261) MiTAC · 2U · AI SERVER / GPU SERVER / HPC / AMD EPYC / LLM INFERENCE
The MiTAC B8261T85E8HR-2T-N-HE (TN85-B8261) is a 2U rack-mount AMD EPYC GPU server supporting dual AMD EPYC 9005/9004 series processors on Socket SP5 with 24 DDR5 DIMM slots up to 6400 MT/s. It accomm
- MSI CG480-S5063 (S5063G480RAE20) MSI · 4U · AI SERVER / GPU SERVER / LLM TRAINING / INFERENCE / NVIDIA MGX
The MSI CG480-S5063 (S5063G480RAE20) is a 4U GPU server on the NVIDIA MGX modular architecture, accepting eight NVIDIA RTX PRO 6000 Blackwell Server Edition professional GPUs at 30°C ambient. It runs
- MSI CG480-S6053 (S6053G480RAU8) MSI · 4U · AI SERVER / GPU SERVER / LLM TRAINING / INFERENCE / NVIDIA MGX
The MSI CG480-S6053 (S6053G480RAU8) is a 4U NVIDIA MGX GPU server running dual AMD EPYC 9005 series processors up to 500W TDP on Socket SP5, with 24 DDR5 DIMM slots in a 12-channel 1DPC configuration
- MSI G4201RAS8U4-P5-Ti MSI · 4U · AI INFERENCE / LLM FINE-TUNING / GPU SERVER / HPC / CLOUD COMPUTING
The MSI G4201RAS8U4-P5-Ti is a 4U rack-mount GPU server with dual 4th/5th Generation Intel Xeon Scalable processors up to 350W TDP on Socket LGA4677 and 32 DDR5 DIMM slots in a 2DPC configuration. It
