Cisco Unveils 102.4 Tbps Silicon One G300 at Cisco Live Amsterdam

Cisco unveiled the Silicon One G300, a 102.4 Tbps switching silicon designed for massive AI cluster buildouts, at Cisco Live EMEA in Amsterdam on February 10, 2026. The breakthrough networking chip sets a new standard for AI backend networking and is engineered to power massive, distributed AI clusters with high performance, security, and reliability.

Intelligent Collective Networking

The G300 uniquely offers Intelligent Collective Networking, which combines an industry-leading fully shared packet buffer, path-based load balancing, and proactive network telemetry. This technology efficiently absorbs bursty AI traffic, responds faster to link failures, and prevents packet drops that can stall compute jobs. With Intelligent Collective Networking, Cisco can deliver 33% increased network utilization and a 28% reduction in job completion time versus simulated non-optimized path selection.

Energy Efficiency and Systems Integration

The Cisco N9000 and Cisco 8000 102.4T Systems powered by Silicon One G300 deliver a new standard in data center performance with both liquid-cooled and air-cooled designs available. The 100% liquid cooled systems achieve a nearly 70% energy efficiency improvement over traditional cooling approaches, addressing the significant power consumption challenges of large-scale AI infrastructure.

Availability

Cisco's G300 chips, systems, and optics are slated to begin shipping later this year. The announcement positions Cisco to compete directly with Broadcom and Nvidia in the rapidly growing AI networking infrastructure market, where demand for high-bandwidth, low-latency networking continues to accelerate as organizations scale AI training and inference workloads.