Gigabyte G593-ZD1-AAX1 8-GPU H200 SXM5 Server SKU

The Gigabyte G593-ZD1-AAX1 is a purpose-built 8-GPU AI server platform engineered around NVIDIA H200 SXM5 accelerators, delivering one of the highest aggregate GPU memory capacities available in a single 5U chassis. Each H200 SXM5 GPU integrates 141 GB of HBM3e memory with a memory bandwidth of 4.8 TB/s per GPU, yielding a combined 1,128 GB of high-bandwidth memory across the eight-accelerator configuration. The system is architected to meet the demands of large-scale generative AI training, deep learning inference, and high-performance computing workloads where memory capacity and inter-GPU bandwidth are the primary bottlenecks.

The G593-ZD1-AAX1 leverages the NVLink high-speed interconnect fabric native to the SXM5 form factor, enabling tight coupling between all eight GPUs for collective operations critical to transformer model training and multi-node GPU cluster deployments. The platform supports dual Intel Xeon Scalable or AMD EPYC host processors depending on board configuration, paired with high-capacity DDR5 system memory and multiple PCIe Gen 5 slots to accommodate high-bandwidth networking adapters such as NVIDIA ConnectX-7 or Quantum-2 InfiniBand HCAs. Comprehensive redundant power and cooling subsystems are integrated to sustain continuous operation at full GPU TDP across all eight accelerators.

Designed for enterprise datacenters and AI infrastructure operators, the G593-ZD1-AAX1 fits within standard 19-inch rack enclosures and supports industry-standard management interfaces including IPMI and Redfish-compliant BMC remote management. The server is targeted at organizations building sovereign AI infrastructure, national research computing facilities, cloud AI service providers, and enterprise ML engineering teams that require dense, thermally managed GPU compute at scale. Its SXM5 mechanical and electrical standard ensures forward compatibility with NVIDIA's GPU computing ecosystem and NVLink Switch fabric expansion.

Ideal for

• Large language model (LLM) pre-training and fine-tuning requiring high aggregate GPU memory to fit multi-hundred-billion parameter models within a single node
• Generative AI inference serving for latency-sensitive enterprise applications where high HBM3e bandwidth reduces time-to-first-token at scale
• Scientific high-performance computing simulations including computational fluid dynamics, molecular dynamics, and climate modeling leveraging FP64 tensor throughput
• Multi-modal foundation model development combining vision, language, and speech modalities that demand large unified GPU memory pools
• Enterprise AI datacenter node consolidation, replacing multiple lower-density GPU nodes with a single high-density 8-GPU chassis to reduce rack footprint and power distribution complexity
• Sovereign AI and national research infrastructure deployments requiring certified, traceable server hardware with enterprise-grade remote management and operational continuity features

Technical specifications

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