Trending Topics
Trending
Cisco announces new networking silicon and optics for AI infrastructure
Cisco has announced new networking silicon, optical technologies and management systems for AI infrastructure.
“AI is redefining what it means to scale infrastructure. It’s no longer just about individual data centers or clusters, but how we connect compute at massive scale without compromising performance or resilience,” Cisco data centre and internet infrastructure senior VP and general manager Kevin Wollenweber said.
Some of the specific key announcements include the production deployments of its 51.2Tbps Silicon One P200 routing systems by Microsoft and Alibaba. The latter company is using the silicon in its eCORE backbone network. Cisco said the architecture is designed to address network congestion, latency and recovery requirements when accelerators at different locations participate in the same workload.
High-performance silicon and optical partnerships
The networking giant has also teamed up with Nvidia to develop an Ultra-Dense Small Form Factor Pluggable (USFP) optical module, which it is said is capable of supporting 400Gbps per lane.
The companies expect the module to provide four times the bandwidth density and 1.3 times the power efficiency of 200Gbps-per-lane optics. They plan to develop an open industry specification through a community-led multi-source project.
“Through our collaboration, USFP is designed to deliver a new level of bandwidth density while preserving the serviceability, deployment flexibility and operational simplicity of pluggable optics,” Cisco optical systems and optics senior vice president and general manager Bill Gartner said.
“By taking a full-stack approach – from silicon and systems to optics and the network operating system – we can optimise every layer to deliver the performance and scale the next generation of AI networks will require.”
Next-generation switching and resilient management
Cisco has also announced the Silicon One G302, a 51.2Tbps switching chip with 200G SerDes designed for top-of-rack deployments. The chip supports 1.6T optics and direct-to-chip liquid cooling. Customer shipments are planned before the end of 2026.
Additionally, it noted its 102.4Tbps Silicon One G300, announced in February, is available for customer sampling, with the first systems shipping to customers.
Cisco has also introduced the 8001, a management system powered by its Silicon One A100 silicon and available in 10G and 1G configurations.
According to Cisco, the system is designed for out-of-band management networks, which provide remote access when production networks fail.
“Operators can use familiar APIs, automation and operating practices from the fabric to the rack, without building and maintaining a separate management environment. Customers can now deploy a simpler, programmable management network that scales with AI infrastructure, and remains available when it is needed most,” the company said.
More from Aimee Chanthadavong
About The Author
Aimee Chanthadavong
NEXT UP
