Brand: AMD | Category: GPUs
SKU: 100-300000075 | Part #: 100-300000075 | MPN: 100-300000075
Contact for Pricing — Request a Quote
The AMD Instinct MI325X GPU Accelerator (part number 100-300000075) delivers 1.458 PetaFLOPS of peak FP32 performance, positioning it as a high-tier compute engine for modern AI and HPC workloads. Paired with 192 GB of HBM3 memory and a memory bandwidth of 3.6 TB/s, this full-height, full-length server card provides the throughput and capacity required for large-scale model training and inference. The MI325X connects via PCIe 5.0 at 128 GB/s for host communication and supports 200 GB/s Infinity Fabric Link connectivity for GPU-to-GPU scaling, enabling multi-accelerator configurations. Peak FP64 performance reaches 729 TFLOPS for double-precision HPC codes, while the card operates within a 420 W power envelope using industry-standard 12V-2x6 / 16-pin server-class power connectors.
AI infrastructure teams, data center architects, and enterprise procurement departments select the AMD Instinct MI325X for large language model training, generative AI inference, and high-performance computing clusters where memory bandwidth and compute density directly impact time-to-solution. AMD's proven enterprise ecosystem and the inclusion of a 3-year manufacturer warranty (with extended options available) reduce deployment risk for mission-critical systems. Operating across a 0–55°C temperature range, the MI325X integrates into existing server cooling infrastructure without exotic environmental requirements, making it suitable for both dedicated AI data centers and hybrid HPC environments.
Organizations evaluating the AMD Instinct MI325X GPU Accelerator 100-300000075 for production deployment should contact Omnixon Global to discuss your configuration requirements, volume pricing, and delivery timelines. Submit a formal RFQ through our team to explore how this accelerator aligns with your infrastructure roadmap.
| Brand | AMD |
| Category | GPUs |
| SKU | 100-300000075 |
| Part Number | 100-300000075 |
| Condition | New |
| GPU Model | MI325X |
| Power Connector | 12V-2x6 / 16-pin (server-class) |
| Warranty | 3-year manufacturer warranty (extended available) |
| Manufacturer Part Number | 100-300000075 |
| Memory | 192 GB HBM3 |
| Peak FP32 Performance | 1.458 PetaFLOPS |
| Peak FP64 Performance | 729 TFLOPS |
| Memory Bandwidth | 3.6 TB/s HBM3 |
| PCIe Interface | PCIe 5.0 (128 GB/s) |
| Infinity Fabric Link | 200 GB/s (GPU-to-GPU) |
| Power Consumption | 420 W |
| Form Factor | Full-height, full-length server card |
| Operating Temperature | 0–55°C |
| Use Cases | LLM training, inference, HPC, generative AI |
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 authorised-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 authorised channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.