Cisco Introduces Silicon One G300 to Boost AI Data Center Performance

Cisco Systems, Inc. announced in a press release the launch of its new Silicon One G300 chip, designed to power and scale AI data centers. The chip supports 102.4 Tbps switching capacity and is built to improve job completion times by 28%, enabling gigawatt-scale AI clusters for training, inference, and real-time workloads.

The G300 powers new Cisco N9000 and 8000 systems, which deliver hyperscale-level performance and reliability for enterprises, service providers, and cloud operators. These systems are available in a fully liquid-cooled design that, combined with new optical components, can enhance energy efficiency by nearly 70%.

Cisco also introduced updates to its Nexus One management platform, providing a unified control plane for faster deployment, predictable scaling, and secure operation of AI networks. The new hardware and software suite aims to simplify operations and lower operating costs for organizations building large-scale AI infrastructure.