NVIDIA B200 SXM6 192GB HBM3e Datacenter GPU

The NVIDIA B200 SXM6 is a flagship datacenter GPU built on the NVIDIA Blackwell architecture, designed to deliver breakthrough performance for the most demanding AI training, large-scale inference, and high-performance computing workloads. Featuring 192 GB of HBM3e memory across a massive 8192-bit memory bus, the B200 SXM6 provides 8 TB/s of memory bandwidth, enabling the acceleration of trillion-parameter foundation models and complex scientific simulations at unprecedented scale. The SXM6 form factor ensures tight integration with NVLink 5 interconnects and NVSwitch fabric within NVIDIA GB200 NVL-based systems, delivering the high-bandwidth, low-latency GPU-to-GPU communication that modern AI infrastructure demands.

The B200 SXM6 incorporates NVIDIA's fifth-generation NVLink with 1.8 TB/s bidirectional bandwidth per GPU, enabling seamless multi-GPU scaling across large server clusters. The chip integrates second-generation Transformer Engine with FP4 and FP8 precision support alongside TF32, BF16, FP16, FP64, and INT8, allowing flexible precision selection that maximizes throughput without sacrificing accuracy for AI model training and inference pipelines. With a TDP of 1000 W operating within the NVLink domain of GB200 systems, the B200 is engineered for liquid-cooled datacenter environments that prioritize sustained peak compute density.

For enterprise IT and cloud infrastructure operators, the NVIDIA B200 SXM6 represents a purpose-built solution for generative AI platforms, LLM serving at scale, autonomous vehicle pipeline development, drug discovery simulations, and climate modeling. Its integration into NVIDIA's Blackwell compute platform — supported by the CUDA ecosystem, NVIDIA AI Enterprise software suite, and cuDNN, NCCL, and TensorRT libraries — provides a mature, optimized software stack that enterprises can deploy across on-premises and hybrid cloud environments. Omnixon Global makes this GPU accessible to enterprise buyers across the UAE, GCC, EMEA, and APAC regions.

Ideal for

• Training and fine-tuning large language models and multimodal foundation models with hundreds of billions to trillions of parameters
• High-throughput generative AI inference serving for enterprise applications including real-time copilots, content generation, and retrieval-augmented generation pipelines
• Accelerated drug discovery and molecular dynamics simulation requiring FP64 double-precision compute at datacenter scale
• Climate modeling, weather forecasting, and large-scale scientific HPC simulations leveraging the full 192 GB HBM3e memory capacity
• Autonomous vehicle and robotics perception model training across large multi-sensor datasets requiring sustained high memory bandwidth
• Enterprise AI platform consolidation within liquid-cooled hyperscale datacenter racks, enabling dense GPU compute nodes with NVLink multi-GPU fabric

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.