Brand: NVIDIA | Category: Servers
SKU: OVX-B200-REF | Part #: OVX-B200-REF | MPN: OVX-B200-REF
Contact for Pricing — Request a Quote
The NVIDIA OVX Server Reference Platform (Blackwell), part number OVX-B200-REF, features integrated NVIDIA ConnectX-7 and BlueField-3 DPU network adapters that deliver high-speed, low-latency connectivity for demanding enterprise workloads. The platform employs fifth-generation NVLink technology with 1.8 TB/s bidirectional bandwidth per GPU, enabling seamless multi-GPU communication at scale. Supporting NVIDIA's modular MGX server architecture in multi-U rack-mount form factor, this reference design accommodates flexible OEM customization while maintaining compatibility with leading Linux distributions—Ubuntu Linux and Red Hat Enterprise Linux (RHEL)—as well as VMware vSphere environments with GPU passthrough capabilities.
Built on the NVIDIA Blackwell architecture, the OVX-B200-REF integrates NVIDIA B200 Tensor Core GPUs with 192 GB of HBM3e memory per GPU and 8 TB/s memory bandwidth per GPU, delivering exceptional throughput for both inference and training workloads. AI infrastructure teams and IT procurement professionals will benefit from native support for NVIDIA AI Enterprise, CUDA, Base Command Manager, NIM microservices, and Omniverse, enabling rapid deployment of generative AI, digital twin simulation, autonomous systems, and high-performance computing applications. Hardware-level confidential computing features are supported via Blackwell architecture security capabilities. Available through Omnixon Global across UAE, GCC, EMEA, and APAC regions, this platform represents NVIDIA's commitment to modular, scalable AI infrastructure. Contact Omnixon Global to request a quotation and explore deployment options tailored to your organization's needs.
| Brand | NVIDIA |
| Category | Servers |
| SKU | OVX-B200-REF |
| Part Number | OVX-B200-REF |
| Condition | New |
| Manufacturer Part Number | OVX-B200-REF |
| Platform Architecture | NVIDIA Blackwell |
| GPU Model | NVIDIA B200 Tensor Core GPU |
| GPU Memory | 192 GB HBM3e per GPU |
| GPU Memory Bandwidth | 8 TB/s per GPU |
| FP8 Tensor Core Performance | 9 PFLOPS per GPU |
| NVLink Generation | Fifth-generation NVLink |
| NVLink Bandwidth | 1.8 TB/s bidirectional per GPU |
| Interconnect Fabric | NVLink Switch with NVSwitch integration |
| Network Adapters | NVIDIA ConnectX-7 / BlueField-3 DPU (configuration-dependent) |
| Server Architecture | NVIDIA MGX modular reference architecture |
| Form Factor | Multi-U rack-mount (reference configuration; exact U-height varies by OEM build) |
| Software Ecosystem | NVIDIA AI Enterprise, CUDA, Base Command Manager, NIM microservices, Omniverse |
| Operating System Support | Ubuntu Linux, Red Hat Enterprise Linux (RHEL), VMware vSphere with GPU passthrough |
| Confidential Computing | Supported via Blackwell architecture hardware security features |
| Target Workloads | Generative AI inference and training, digital twin simulation, autonomous systems, HPC |
| Regional Availability | UAE, GCC, EMEA, APAC via Omnixon Global |
Yes in Dubai, and on quotation for the broader GCC. Includes rack-and-stack, cable management, OS install, and post-install verification.
