Brand: Supermicro | Category: GPUs
SKU: 1-300-100-006 | Part #: 1-300-100-006 | MPN: 1-300-100-006
Contact for Pricing — Request a Quote
The Intel Gaudi 3 OAM 128GB HBM2E (8-module system card), offered by Supermicro under manufacturer part number 1-300-100-006, is a high-density AI accelerator system card integrating eight Intel Gaudi 3 OAM (Open Accelerator Module) processors onto a single board. Built on Intel's third-generation Gaudi architecture, each OAM die features dedicated matrix multiplication engines and tensor cores optimized for large-scale deep learning training and inference, with the eight-module configuration delivering a combined 128GB of HBM2E memory with high aggregate memory bandwidth across the full card. The Gaudi 3 architecture provides significant generational improvements in FP8, BF16, and FP16 compute throughput compared to its predecessors, making it well suited for the demanding matrix operations at the heart of large language model (LLM) training and transformer-based inference workloads.
The OAM form factor adheres to the OAM Universal Baseboard (UBB) open standard, enabling standardized integration into compatible OAM-ready server platforms including Supermicro's purpose-built AI training servers. Each Gaudi 3 OAM module incorporates onboard 24 x 200GbE RoCEv2 high-speed interconnects (across the full eight-module card), enabling direct, low-latency scale-out networking between nodes without requiring a separate, proprietary switching fabric. This design reduces infrastructure complexity for multi-node AI clusters and allows the card to participate in large distributed training runs across hundreds of accelerators using standard Ethernet-based RDMA protocols.
As a Supermicro-integrated system card, the 8-module Gaudi 3 configuration is intended for enterprise datacenter deployments where maximum accelerator density, memory capacity, and scale-out networking are critical. It is supported by Intel's open software stack, including the Habana SynapseAI SDK, which provides compatibility with PyTorch and TensorFlow frameworks, enabling enterprise teams to migrate and optimize existing AI workflows without dependency on proprietary toolchains. This accelerator is available through Omnixon Global for enterprise buyers across the UAE, GCC, EMEA, and APAC regions.
| Manufacturer | Supermicro |
| Manufacturer Part Number | 1-300-100-006 |
| Product Name | Intel Gaudi 3 OAM 128GB HBM2E (8-module system card) |
| Accelerator Architecture | Intel Gaudi 3 (third-generation Gaudi) |
| Number of OAM Modules | 8 |
| Form Factor | OAM Universal Baseboard (UBB) system card |
| Total HBM2E Memory | 128 GB (across 8 modules) |
| Memory per Module | 16 GB HBM2E |
| Memory Type | HBM2E |
| Supported Precisions | FP8, BF16, FP16, FP32 |
| Scale-Out Networking | 24 x 200GbE RoCEv2 integrated ports (across 8-module card) |
| Networking Protocol | RoCEv2 (RDMA over Converged Ethernet) |
| Software Stack | Intel Habana SynapseAI SDK |
| Framework Support | PyTorch, TensorFlow |
| Target Platform | Supermicro OAM-compatible AI training servers |
| Use Case Classification | AI training, large-scale inference, HPC |
| Compliance | OAM Open Standard (Open Accelerator Module) |
Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.
| Brand | Supermicro |
| Category | GPUs |
| SKU | 1-300-100-006 |
| Part Number | 1-300-100-006 |
| Condition | New |
| Manufacturer Part Number | 1-300-100-006 |
| Product Name | Intel Gaudi 3 OAM 128GB HBM2E (8-module system card) |
| Accelerator Architecture | Intel Gaudi 3 (third-generation Gaudi) |
| Number of OAM Modules | 8 |
| Form Factor | OAM Universal Baseboard (UBB) system card |
| Total HBM2E Memory | 128 GB (across 8 modules) |
| Memory per Module | 16 GB HBM2E |
| Memory Type | HBM2E |
| Supported Precisions | FP8, BF16, FP16, FP32 |
| Scale-Out Networking | 24 x 200GbE RoCEv2 integrated ports (across 8-module card) |
| Networking Protocol | RoCEv2 (RDMA over Converged Ethernet) |
| Software Stack | Intel Habana SynapseAI SDK |
| Framework Support | PyTorch, TensorFlow |
| Target Platform | Supermicro OAM-compatible AI training servers |
| Use Case Classification | AI training, large-scale inference, HPC |
| Compliance | OAM Open Standard (Open Accelerator Module) |
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.
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.
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.
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.