Cisco Nexus 9364C-GX 64-port 100/400G Switch

The Cisco Nexus 9364C-GX is a high-density, fixed-form-factor spine-layer switch engineered for next-generation data center and cloud fabrics demanding 400 Gigabit Ethernet throughput. Built on the Cisco Cloud Scale ASIC (Cloudscale), the platform delivers a switching capacity of 25.6 Tbps with support for 64 QSFP-DD ports that operate natively at 400G, while retaining backward compatibility with 40G, 100G, and breakout configurations down to 10G. The switch occupies 2U of rack space and is purpose-designed for leaf-spine architectures where east-west traffic volumes are driven by distributed AI/ML training clusters, high-frequency analytics pipelines, and large-scale virtualization environments.

The Nexus 9364C-GX runs Cisco NX-OS in standalone mode and is also fully supported within Cisco Application Centric Infrastructure (ACI) as a spine node, enabling policy-driven automation and microsegmentation across multi-tenant fabrics. The platform supports advanced buffer management and features Virtual Output Queuing (VOQ) to minimize head-of-line blocking under congested, many-to-one traffic patterns common in GPU-to-GPU workloads. Telemetry capabilities include streaming telemetry via gRPC/gNMI and support for In-band Network Telemetry (INT), giving operations teams real-time, per-flow visibility at wire rate without impacting forwarding performance.

Designed for operational continuity in mission-critical environments, the 9364C-GX incorporates redundant, hot-swappable power supplies and fan modules with front-to-rear and rear-to-front airflow options to accommodate varied data center cooling architectures. The switch integrates with Cisco Nexus Dashboard and supports open standards including OpenConfig, YANG data models, and RESTCONF/NETCONF APIs, enabling seamless integration into NetDevOps workflows and third-party orchestration platforms. Its combination of port density, forwarding scale, and programmability makes it a foundational building block for enterprises modernizing their network infrastructure for AI/ML, NVMe-oF storage, and hyperscale workloads.

Ideal for

• Spine-layer node in leaf-spine fabrics supporting GPU clusters for distributed AI and machine learning model training with high-bandwidth, low-latency east-west interconnects
• High-density 400G spine aggregation for large-scale hyperconverged infrastructure and private cloud deployments requiring non-blocking throughput across thousands of server endpoints
• NVMe-over-Fabrics (NVMe-oF) storage networking backbone where consistent low latency and deep buffer management are critical for all-flash array performance
• Core switching layer in Cisco ACI multi-pod or multi-site fabrics, providing policy enforcement and automated provisioning across geographically distributed data center campuses
• Network telemetry aggregation point in real-time analytics environments, leveraging streaming telemetry and INT to feed observability platforms with per-flow, per-queue metrics at wire rate
• Interconnect fabric for high-performance computing (HPC) clusters in financial services, genomics, or scientific research environments requiring deterministic throughput across dense server pools

Technical specifications

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