Available In Stock
Quantum-2 InfiniBand
Fabric Switch
Ultra-low latency networking platform for hyperscale AI fabrics, high-density clusters and unified GPU communication.
Architecture
Quantum 2
Process
400G NDR
Transistors
51.2Tb/s
Memory
64 Ports
Technical Infrastructure
Comprehensive performance metrics and architectural specifications.
| Parameter | Value |
|---|---|
| Brand | NVIDIA Networking |
| Application | AI Cluster Fabric |
| Ports | 64 x 400Gb/s |
| Switching Capacity | 51.2Tb/s |
| Latency | 130ns |
| Rack Unit | 1U |
| Cooling | Redundant hot-swap fans |
| Power | Dual redundant PSU |
| Management | Web / CLI / Telemetry |
| Fabric | NDR InfiniBand |
| Deployment | GPU Pods and Clusters |
| Use Case | Scale-out AI networking |
*Switching capacity and latency vary by deployment topology.
View Detailed Architecture Whitepaper open_in_newPrecision Engineering
Explore every detail of the hardware that powers the modern AI era.
Fabric Ports
Dense 400G connectivity for large training clusters.
Cluster Cabling
Low-latency topology for synchronized AI workloads.
Rack Integration
Purpose-built for modern pod-level deployment standards.
Ready to Transform Your Infrastructure?
Our engineering team is ready to assist you in configuring the ideal deployment for your specific workload.