Brand: Supermicro | Category: Servers
SKU: SUPE-SRSGB200NVL7201 | Part #: SRS-GB200NVL72-01 | MPN: SRS-GB200NVL72-01
Contact for Pricing — Request a Quote
The Supermicro GB200 NVL72 Rack-Scale System (SRS-GB200NVL72) from Supermicro is enterprise-grade Servers hardware built for data centers. Rack-scale Grace Blackwell NVL72 system integrating 72 Blackwell GPUs and 36 Grace CPUs with NVLink fabric for exascale AI factory deployments. Available in brand new condition. Available from Omnixon Global — contact our team to request a quote.
| Brand | Supermicro |
| Category | Servers |
| SKU | SUPE-SRSGB200NVL7201 |
| Part Number | SRS-GB200NVL72-01 |
| Condition | New |
| Product Line | Grace Blackwell Rack-Scale System |
| Model | SRS-GB200NVL72 |
| GPU Architecture | NVIDIA Blackwell (B200) |
| Total GPUs per Rack | 72x NVIDIA B200 GPUs |
| CPU Architecture | NVIDIA Grace (ARM Neoverse V2) |
| Total CPUs per Rack | 36x NVIDIA Grace CPUs (paired 2:1 GPU-to-CPU via NVLink C2C) |
| GPU Interconnect | NVLink 5th Generation, all-to-all non-blocking fabric across all 72 GPUs |
| NVLink Bandwidth per GPU | 1.8 TB/s bidirectional |
| Grace-to-Blackwell C2C Bandwidth | 900 GB/s bidirectional per GB200 Superchip pair |
| Peak AI Compute (FP8) | >11.5 ExaFLOPS aggregate across 72 B200 GPUs |
| Total HBM3e Capacity | ~13.5 TB aggregate (192 GB HBM3e per B200 GPU) |
| HBM3e Memory Bandwidth per GPU | 8 TB/s |
| Grace CPU Memory | LPDDR5X with 480 GB/s bandwidth per Grace CPU; 128 GB per CPU module |
| Thermal Design | Direct liquid cooling (DLC) with manifold distribution; supports >120 kW rack-level thermal load |
| Rack Form Factor | Full rack (42U-class rack-scale unit), factory-integrated |
| Network Fabric Support | InfiniBand NDR400 and/or 400GbE per compute tray for scale-out interconnect |
| Storage Interface | NVMe-oF / GPUDirect Storage compatible; external high-performance storage fabric integration |
| System Management | IPMI 2.0, Redfish API, Supermicro BMC, NVIDIA DOCA integration |
| Software Ecosystem | NVIDIA AI Enterprise, CUDA 12.x, NVLink Switch System software, NGC container registry compatible |
| Power Supply | Redundant high-efficiency 277V AC or 380V DC power shelf; N+N redundancy configuration |
| Target Deployment | Hyperscale AI data centers, enterprise AI factories, national AI supercomputing facilities |
| Launch | 2025 Q3 |
The Supermicro GB200 NVL72 Rack-Scale System (SRS-GB200NVL72) is built for enterprise data-center workloads — virtualization (VMware, Proxmox, Nutanix), private cloud, database hosting, and AI/ML training. It fits standard EIA-310 server racks and supports redundant PSUs and hot-swap drives common in production environments.
Key specifications for the Supermicro GB200 NVL72 Rack-Scale System (SRS-GB200NVL72): new condition; manufacturer Supermicro; product line Grace Blackwell Rack-Scale System; model SRS-GB200NVL72; gpu architecture NVIDIA Blackwell (B200); total gpus per rack 72x NVIDIA B200 GPUs; cpu architecture NVIDIA Grace (ARM Neoverse V2). Manufacturer part number SRS-GB200NVL72-01. For the full datasheet with electrical, environmental, and compliance details, contact our pre-sales engineering team.