Cisco Systems has introduced two new networking chips, the Silicon One G200 and G202, designed for web-scale spine roles and Ethernet-based artificial intelligence/machine learning (AI/ML) deployments.
The G200 is a 5 nm, 51.2 Tbps device, while the G202 is a 5 nm, 25.6 Tbps device. Both chips are designed to be highly programmable and efficient, and they offer a unified architecture that can be used for a variety of networking applications.
Cisco believes that the G200 and G202 will be the most powerful networking chips on the market for AI and ML workloads. The chips offer double the performance of the previous generation of Silicon One chips, and they can connect up to 32,000 graphics processing units (GPUs) together.
The G200 and G202 are currently being tested by five out of the six major cloud providers, and Cisco expects them to be widely adopted in the coming years. The chips are well-positioned to meet the growing demand for AI and ML applications, and they will help to accelerate the adoption of these technologies.
The sources for this piece include an article in Reuters.