Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack
Brand: Gigabyte | Category: GPUs
SKU: G593-SE0-AAX1 | Part #: G593-SE0-AAX1 | MPN: G593-SE0-AAX1
Contact for Pricing — Request a Quote
About the Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack
The Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack (part number G593-SE0-AAX1) delivers massive parallel compute density in a full-rack Open Rack v3 form factor, housing 72 x NVIDIA B200 GPU dies organized as 36 x NVIDIA GB200 Supercbips—each combining 2 B200 GPUs and 1 NVIDIA Grace CPU (Arm Neoverse V2-based). The unified architecture pools 13.5 TB of HBM3e memory across the entire rack, with each GPU delivering up to 1.4 TB/s memory bandwidth. Fifth-generation NVLink interconnect unifies all 72 GPUs into a single domain, providing up to 1.8 TB/s bisection bandwidth for coherent multi-GPU workloads without network bottlenecks.
Gigabyte's liquid cooling architecture—combining direct liquid cooling and rear-door heat exchanger design—ensures thermal efficiency across maximum GPU utilization. The system achieves up to 1.4 exaflops AI inference performance (FP4) and up to 720 petaflops AI training performance (FP8), making it purpose-built for large-language model training, generative AI, and high-performance computing at scale. Network connectivity via NVIDIA Quantum-2 InfiniBand or Spectrum-X Ethernet enables multi-rack clustering, while Linux operating system support (Ubuntu, Red Hat Enterprise Linux) and Baseboard Management Controller (BMC) with IPMI/Redfish integration streamline deployment and remote management. AI infrastructure teams and hyperscale operators will find this platform essential for next-generation transformer and foundation model workloads. Contact Omnixon Global to discuss your AI infrastructure requirements and request an RFQ for the Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack.
Typical Deployment Scenarios
- Large language model (LLM) training and fine-tuning at scale
- Generative AI inference serving and real-time model deployment
- High-performance computing (HPC) simulation and scientific research
- Multi-modal foundation model development and optimization
- Enterprise data center AI acceleration and cloud service delivery
Technical Specifications
| Brand | Gigabyte |
| Category | GPUs |
| SKU | G593-SE0-AAX1 |
| Part Number | G593-SE0-AAX1 |
| Condition | New |
| Manufacturer Part Number | G593-SE0-AAX1 |
| Product Name | Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack |
| GPU Platform | NVIDIA GB200 NVL72 |
| GPU Architecture | NVIDIA Blackwell |
| Total GPUs per Rack | 72 x NVIDIA B200 GPU dies |
| CPU Configuration | 36 x NVIDIA Grace CPUs (Arm Neoverse V2-based) |
| Compute Module Type | NVIDIA GB200 Superchip (2x B200 GPU + 1x Grace CPU per Superchip) |
| AI Inference Performance (FP4) | Up to 1.4 exaflops per rack |
| AI Training Performance (FP8) | Up to 720 petaflops per rack |
| Total GPU HBM3e Memory | 13.5 TB per rack (72 x 192 GB HBM3e) |
| GPU Memory Bandwidth | Up to 1.4 TB/s per GPU (HBM3e) |
| NVLink Interconnect | Fifth-generation NVLink, all 72 GPUs in a unified domain |
| NVLink Total Bandwidth | Up to 1.8 TB/s bisection bandwidth across the NVLink domain |
| Cooling Architecture | Liquid cooling (direct liquid cooling / rear-door heat exchanger) |
| Form Factor | Full rack (Open Rack v3 compatible) |
| Network Support | NVIDIA Quantum-2 InfiniBand / Spectrum-X Ethernet |
| Operating System Support | Linux (Ubuntu, Red Hat Enterprise Linux) |
| Management | Baseboard Management Controller (BMC) with IPMI/Redfish support |
Frequently Asked Questions about Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack
What server platforms accept the Gigabyte G593-SE0 NVIDIA GB200 NVL72 AI Server Rack?
Reference servers include Dell PowerEdge XE9680 / XE9712, HPE Cray XD670, Lenovo ThinkSystem SR685a / SR675 V3, Supermicro AS-A21GE / SYS-821GE, Gigabyte G593 / G894, ASUS ESC. Share your target platform in the RFQ and we will confirm chassis-to-GPU compatibility and recommended NIC pairing.
How long is the lead time on AI GPUs?
Highly model-dependent. L40S / RTX-class: typically 3-6 weeks. H100/H200/B200 in SXM form factor: 12-16 weeks for whole-platform allocations. We quote genuine-channel ETAs only — no grey-market promises.
Do you supply matched networking (Quantum InfiniBand / Spectrum-X)?
Yes — Omnixon stocks the full NVIDIA networking lineup (Quantum-2 / Quantum-X InfiniBand, Spectrum-X Ethernet, ConnectX NICs, BlueField DPUs) so we can quote a complete training-cluster BOM, not just the GPUs.
Can you help with NVIDIA AI Enterprise licensing?
Yes. We hold genuine-channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.