In the era of AI, a unit of compute is no longer measured by just the number of servers. Interconnected GPUs, CPUs, memory, storage, and these resources across multiple nodes in racks construct today's artificial Intelligence. The infrastructure requires high-speed and low-latency network fabrics, and carefully designed cooling technologies and power delivery to sustain optimal performance and efficiency for each data center environment. Supermicro’s SuperCluster solution provides end-to-end AI data center solutions for rapidly evolving Generative AI and Large Language Models (LLMs).
Complete Integration at Scale
Design and build of full racks and clusters with a global manufacturing capacity of up to 5,000 racks per month
Test, Validate, Deploy with On-site Service
Proven L11, L12 testing processes thoroughly validate the operational effectiveness and efficiency before shipping
Liquid Cooling/Air Cooling
Fully integrated liquid-cooling or air cooling solution with GPU & CPU cold plates, Cooling Distribution Units and Manifolds
Supply and Inventory Management
One-stop-shop to deliver fully integrated racks fast and on-time to reduce time-to-solution for rapid deployment
The full turn-key data center solution accelerates time-to-delivery for mission-critical enterprise use cases, and eliminates the complexity of building a large cluster, which previously was achievable only through the intensive design tuning and time-consuming optimization of supercomputing.
NVIDIA GB200 NVL72
Liquid-cooled Exascale Compute in a Single Rack
- 72x NVIDIA Blackwell B200 GPUs acting as one GPU with a massive pool of HBM3e memory (130TB per rack)
- 9x NVLink Switch, 4 ports per compute tray connecting 72 GPUs to provide 1.8TB/s GPU-to-GPU interconnect
- Supermicro 250kW capacity in-row coolant distribution unit (with) redundant PSU and dual hot-swap pumps, or 240kW/180kW capacity liquid-to-air solution (no facility water required)
- Supermicro’s end-to-end liquid-cooling solution and comprehensive onsite deployment services
- Ready for advanced networking technologies for scaling including NVIDIA BlueField®-3 SuperNIC, Spectrum™-X, Quantum-2, and next generation 800 Gb/s networking platforms
Compute Tray
- ARS-121GL-NBO
Leading Liquid-Cooled AI Cluster
With 32 NVIDIA HGX H100/H200 8-GPU, 4U Liquid-cooled Systems (256 GPUs) in 5 Racks
- Doubling compute density through Supermicro’s custom liquid-cooling solution with up to 40% reduction in electricity cost for data center
- 256 NVIDIA H100/H200 GPUs in one scalable unit
- 20TB of HBM3 with H100 or 36TB of HBM3e with H200 in one scalable unit
- 1:1 networking to each GPU to enable NVIDIA GPUDirect RDMA and Storage for training large language model with up to trillions of parameters
- Customizable AI data pipeline storage fabric with industry leading parallel file system options
- Supports NVIDIA Quantum-2 InfiniBand and Spectrum™-X Ethernet platform
- Certified for NVIDIA AI Enterprise Platform including NVIDIA NIM microservices
Compute Node
Proven Design
With 32 NVIDIA HGX H100/H200 8-GPU, 8U Air-cooled Systems (256 GPUs) in 9 Racks
- Proven industry leading architecture for large scale AI infrastructure deployments
- 256 NVIDIA H100/H200 GPUs in one scalable unit
- 20TB of HBM3 with H100 or 36TB of HBM3e with H200 in one scalable unit
- 1:1 networking to each GPU to enable NVIDIA GPUDirect RDMA and Storage for training large language model with up to trillions of parameters
- Customizable AI data pipeline storage fabric with industry leading parallel file system options
- Supports NVIDIA Quantum-2 InfiniBand and Spectrum-X Ethernet platform
- Certified for NVIDIA AI Enterprise Platform including NVIDIA NIM microservices
Compute Node
Cloud-Scale Inference
With 256 NVIDIA GH200 Grace Hopper Superchips, 1U MGX Systems in 9 Racks
- Unified GPU and CPU memory for cloud-scale high volume, low-latency, and high batch size inference
- 1U Air-cooled NVIDIA MGX Systems in 9 Racks, 256 NVIDIA GH200 Grace Hopper Superchips in one scalable unit
- Up to 144GB of HBM3e + 480GB of LPDDR5X, enough capacity to fit a 70B+ parameter model in one node
- 400Gb/s InfiniBand or Ethernet non-blocking networking connected to spine-leaf network fabric
- Customizable AI data pipeline storage fabric with industry leading parallel file system options
- NVIDIA AI Enterprise Ready including NVIDIA NIM microservices
Compute Node
Enterprise 3D + AI
With 32 4U PCIe GPU Air-Cooled Systems (up to 256 NVIDIA L40S GPUs) in 5 Racks
- Maximize multi-workload performance for enterprise AI-enabled workflows. Optimized for NVIDIA Omniverse with OpenUSD.
- 256 NVIDIA L40S GPUs in one scalable unit
- 12TB of GPU memory and 32TB of system memory in one scalable unit
- Scale-out with 400Gb/s NVIDIA Spectrum™-X Ethernet
- Customizable data storage fabric with industry leading parallel file system options
- Certified for NVIDIA Omniverse Enterprise with included Enterprise Support Services