KR6298-X2 – 6U 8-OAM AI Server with Intel® Habana® Gaudi®2

The KR6298-X2 6U AI Server is purpose-built for next-generation AI, LLM training and inference, and high-performance cloud workloads.
Equipped with 8x Intel® Habana® Gaudi®2 OAM accelerators and dual 4th/5th Gen Intel® Xeon® Scalable CPUs, the KR6298-X2 delivers scalable high-speed compute performance and optimized energy efficiency.

With 96GB HBM3 memory per accelerator, 4.2TB/s converged P2P bandwidth, and 4000Gbps cross-node communication, this platform ensures high-throughput training, inference, and pipeline parallelism for demanding AI environments.

Main Application Areas:

• Generative AI & LLM Workloads: Large-scale model training, fine-tuning, and inference

• Mainstream AI Frameworks: TensorFlow, PyTorch, and AI development tools

• High-Performance Cloud Computing: Multi-tenant AI clusters, HPC pipelines

• Data-Intensive Applications: Video processing, AI pipelines, large-scale analytics

With its cutting-edge memory, interconnect, and efficiency, the KR6298-X2 is the ideal solution for enterprises, research institutions, and data centers developing large-scale AI models and generative AI workloads.

Price on request – daily updates apply. Please check availability before ordering.