The Cisco Nexus N9K-C93360YC-FX2 packs extreme port density for AI workloads with 96 ports of 25 Gigabit Ethernet and 12 ports of 100 Gigabit uplinks. This high-density leaf switch maximizes GPU server connectivity in space-constrained data centers.

Perfect for AI training clusters where every rack unit counts. Connect 96 GPU servers with full 25G bandwidth per node, while 12x 100G uplinks provide non-blocking connectivity to spine switches. The FX2 platform delivers consistent low latency even under heavy AI traffic loads.

Engineered for machine learning operations at scale. Supports RDMA over Converged Ethernet (RoCE) for high-performance GPU-to-GPU communication. Works seamlessly with NVIDIA DGX systems, AWS instances, and custom AI infrastructure.

Key Benefits:

• 2.88 Tbps total bandwidth handles massive datasets
• 96 server connections per 1RU switch saves rack space
• Sub-microsecond latency optimized for AI workloads
• RoCE support accelerates GPU cluster performance
• Cisco Tetration integration for AI traffic analytics

Ships configured and tested. Reef Telecom provides deployment planning and ongoing support.