Now Available
SUPERMICRO GB300 NVL72
Rack-scale AI system from Supermicro with 72 Blackwell Ultra GPUs and 36 Grace CPUs in a single liquid-cooled 48U rack. Up to 21 TB HBM3e GPU memory. Built for large-scale AI training, reasoning, and inference.
Plan Your GB300 Deployment View Full SpecificationsPRODUCT DETAILS
SUPERMICRO SRS-GB300-NVL72
72 BLACKWELL ULTRA GPUS
72 NVIDIA B300 (Blackwell Ultra) GPUs with up to 288 GB HBM3e per GPU and approximately 21 TB total GPU memory. The entire rack operates as a single NVLink domain.
36 NVIDIA GRACE CPUS
36 Arm-based NVIDIA Grace CPUs tightly coupled with the GPU fabric. Up to 17 TB LPDDR5X system memory. Purpose-built to minimize CPU-GPU bottlenecks in AI workloads.
NVLINK 5 AT RACK SCALE
1.8 TB/s NVLink bandwidth per GPU, 130 TB/s aggregate across the rack. All 72 GPUs communicate as a single compute domain, eliminating multi-node overhead for large models.
SUPERMICRO DIRECT LIQUID COOLING
Full rack liquid cooling with in-rack CDU and redundant pumps, backed by Supermicro’s leadership in direct liquid cooling technology. Required for the thermal envelope of 72 GPUs at 132 to 140 kW per rack.
SPECIFICATIONS
Model
Supermicro SRS-GB300-NVL72
System Type
Rack-scale AI system (NVIDIA GB300 NVL72 reference design)
GPU
72x NVIDIA B300 (Blackwell Ultra) GPUs
GPU Memory
Up to 288 GB HBM3e per GPU (approximately 21 TB total)
GPU Interconnect
NVLink 5 with NVSwitch; 1.8 TB/s per GPU; 130 TB/s aggregate
CPU
36x NVIDIA Grace (Arm Neoverse V2), 72 cores per CPU
System Memory
Up to 17 TB LPDDR5X
Compute Trays
18x 1U compute trays (4 GPUs + 2 Grace CPUs per tray)
NVLink Switch Trays
9x NVLink Switch trays
Storage
Up to 144x E1.S PCIe 5.0 drive bays (8 per compute tray)
Networking
ConnectX-8 SuperNIC; 800 Gb/s per GPU; Quantum-X800 InfiniBand or Spectrum-X Ethernet
Cooling
Direct liquid cooling (required); in-rack CDU with redundant pumps
Power
8x 1U 33 kW power shelves; 48V DC busbar
Operating Power
Approximately 132 to 140 kW per rack
Form Factor
48U rack
Starting Price
Contact Us
PLATFORM
NVIDIA GB300 NVL72: RACK-SCALE AI FOR THE REASONING ERA
The GB300 NVL72 is NVIDIA’s flagship rack-scale AI platform, purpose-built for the shift toward test-time scaling and AI reasoning workloads. Compared to Hopper-based systems, the GB300 NVL72 delivers a 10x improvement in tokens per second per user and a 5x improvement in throughput per megawatt, combining to a 50x increase in overall AI factory output.
The Supermicro SRS-GB300-NVL72 is one of the GB300 NVL72 systems available through Arc Compute. Supermicro brings industry-leading direct liquid cooling technology, manufacturing capacity, and end-to-end data center building block solutions to this platform. Our team handles capacity planning, facility assessment, and deployment support for rack-scale and multi-rack installations.
Explore all NVIDIA GB300 NVL72 systems
WHY ARC COMPUTE
THE TEAM BEHIND YOUR
INFRASTRUCTURE
RACK-SCALE DEPLOYMENT EXPERIENCE
Our team has hands-on experience planning and deploying high-density, liquid-cooled GPU infrastructure. Rack-scale systems require rack-scale planning, and we handle every step.
FAST, PREDICTABLE TIMELINES
We work to timelines that match your business reality. Systems are validated and deployed with full support included. planning, and we handle every step.
FACILITY ASSESSMENT AND PLANNING
Power, cooling, floor loading, and CDU integration all need to be validated before a GB300 rack arrives. We work with your facility team to ensure readiness.
LONG-TERM
PARTNERSHIP
We stay engaged after deployment. As your workloads evolve and your infrastructure grows, we help you plan, optimize, and scale.
PLAN YOUR SUPERMICRO GB300 NVL72 DEPLOYMEN
Tell us about your infrastructure goals, facility, and timeline. Our team will help you plan a Supermicro GB300 NVL72 deployment from capacity planning through installation.
Plan Your GB300 Deployment