ASUS ESC8000-E12 with Intel Gaudi 3 Configuration

The ASUS ESC8000-E12 with Intel Gaudi 3 Configuration is a high-density 4U rack server engineered for large-scale AI training and inference workloads. Built on the ASUS ESC8000-E12 platform paired with Intel Gaudi 3 AI accelerators, this system integrates Intel's third-generation Gaudi architecture, delivering high-bandwidth interconnects and substantial HBM2e memory capacity per accelerator. The configuration supports up to eight Intel Gaudi 3 mezzanine accelerator cards, enabling tightly coupled all-to-all communication across accelerators via integrated 24-port 200GbE RoCE networking fabric built directly into each Gaudi 3 die.

At the processor level, the ESC8000-E12 platform supports dual 5th Generation Intel Xeon Scalable processors (Emerald Rapids) with high-speed DDR5 memory across 32 DIMM slots, providing the CPU-side compute and memory bandwidth necessary to feed eight Gaudi 3 accelerators without bottlenecks. The system incorporates PCIe 5.0 connectivity, redundant 3000W platinum-rated power supplies, and enterprise-grade thermal management suited to dense AI compute environments. Storage options include front-accessible NVMe and SATA drive bays, and the system supports OCP 3.0 network mezzanine cards for flexible fabric integration.

This configuration is purpose-built for enterprise data centers deploying generative AI, large language model training, and high-throughput inference at scale. As an Intel Gaudi 3-based platform, it offers an alternative to GPU-based AI infrastructure with a software stack compatible with PyTorch and optimized via Intel's SynapseAI SDK. The system is suited to organizations in financial services, healthcare AI, telecommunications, and cloud service environments across the UAE, GCC, EMEA, and APAC regions seeking high-performance, standards-based AI compute infrastructure.

Ideal for

• Large language model (LLM) pre-training and fine-tuning at enterprise scale using distributed data and model parallelism across eight Gaudi 3 accelerators
• High-throughput generative AI inference for production deployment of transformer-based models in natural language processing and multimodal applications
• Deep learning training for computer vision, recommendation systems, and speech recognition in data center environments requiring dense compute per rack unit
• Financial services AI workloads including risk modeling, fraud detection, and quantitative analytics requiring high-bandwidth accelerator-to-accelerator communication
• Healthcare and life sciences AI research including genomics processing, medical imaging analysis, and drug discovery model training
• Sovereign AI data center deployments across the GCC and EMEA regions where organizations require on-premises, high-performance AI infrastructure with standards-based interconnects

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.