Cisco Silicon One G-Series
The G-Series delivers industry-leading switching performance for AI scale-out and scale-up networking. From the flagship 102.4 Tbps G300 to the versatile G100, these processors power the world's largest AI clusters with deterministic low latency, advanced load balancing, and unmatched telemetry.
Product Overview
102.4 Tbps Flagship
The G300 delivers 102.4 Tbps with 64x 1600GE ports, purpose-built for next-generation AI backend fabrics and web-scale data centers.
AI-Optimized
Advanced congestion-aware load balancing, DLB, packet spraying, and SRv6 uSID for optimal job completion time in massive AI clusters.
Ultra-Low Latency
Deterministic, latency-optimized programmable NPU with run-to-completion flexibility for complex flows and AI workloads.
Unified Packet Buffer
Large, fully shared on-die packet buffer reduces packet loss and minimizes PFC events for reliable GPU data delivery.
Advanced Telemetry
P4-INT, IFA, sFlow, NetFlow, in-band telemetry with temporal dynamics and live network triggers for postmortem analysis.
Programmable
One SDK across all Silicon One processors with SAI support, SONiC reference, and IDE-based application development.
Ready to build with Silicon One G-Series?
Our data center specialists will help you select the right processor for your AI fabric, cloud, or enterprise deployment.
- Volume pricing available
- Cisco SmartNet bundling
- Professional deployment services
- Response within one business day