San Jose – October 12, 2026 -- Cisco and NVIDIA plan to develop an open industry specification for a 400Gb/s-per-lane pluggable optical module that is expected to deliver 4x the bandwidth density of 200Gb/s-per-lane optics. The companies announced the Ultra-Dense Small Form Factor Pluggable (USFP) at the Open Compute Project (OCP) Global Summit 2026.
Cisco also said the first systems powered by its 102.4T Silicon One G300 are shipping to customers. Its 51.2T P200-based systems are running in production at hyperscalers, including Microsoft and Alibaba.
USFP targets 4x bandwidth density and 1.3x power efficiency
Cisco and NVIDIA said USFP is expected to deliver 1.3x the power efficiency of 200Gb/s-per-lane optics. They say it relies on a proven operating model to reach 400Gb/s per lane, and that it preserves serviceability, deployment flexibility and the operational simplicity of today's pluggable optics.
The companies plan to publish the specification through an upcoming community-led multi-source project to drive broad adoption.
Microsoft and Alibaba run Silicon One P200 in production
Cisco said it is the first in the industry to announce 51.2T systems deployed in production by leading hyperscalers. Microsoft said it has moved into production on P200-powered systems. It cited the chip's speed and enhanced buffering for its cloud and high-performance services.
Alibaba Cloud Intelligence Group said it deployed its in-house Quantum Cat router at massive scale on its eCORE backbone network. Each router provides 51.2 Tbps of routing capacity with a fully shared deep-buffer architecture. Alibaba paired the P200 with its custom systems and the SONiC operating system.
Cisco's scale-across architecture combines the P200, deep-buffered systems and coherent optics to link AI fabrics across sites as a single cluster. Cisco says that when accelerators at different sites work on the same job, congestion, latency and recovery directly affect compute utilization and job progress.
G300 ships, G302 adds 51.2T top-of-rack silicon with 1.6T optics
The 102.4T G300, announced earlier this year, is available for customer sampling. Cisco also introduced the Silicon One G302, which it describes as the first 51.2T switching silicon with 200G SerDes designed for top-of-rack deployments. The G302 features Cisco 1.6T optics and direct-to-chip liquid cooling integration. Cisco plans to begin shipping it before the end of 2026.
Cisco is a member of OCP's ESUN workstream. It is working with that group on the 1.0 and future specifications for scale-up workloads.
Live Protect for SONiC arrives in mid-2027
Cisco customers can adopt Cisco-validated SONiC with Nexus One, the company's next-generation data center architecture. Starting mid-year 2027, Cisco will extend that support with Live Protect. Using Cisco runtime technology, Live Protect is designed to reduce exposure between a vulnerability disclosure and a patch on supported Cisco products and releases. It gives SONiC customers time to test and deploy the permanent fix without disrupting operations.
Cisco also launched the Cisco 8001, an out-of-band management system powered by Silicon One A100 and available as 10G or 1G. It brings SONiC to the management network so operators can use the same APIs and automation from fabric to rack. Cisco says this removes the need for a separate management environment for thousands of GPU nodes.
Kevin Wollenweber, SVP and General Manager of Data Center and Internet Infrastructure at Cisco, said the challenge is to connect compute "at massive scale without compromising performance or resilience."
Cisco says Silicon One, introduced in 2019, now scales from 100G to 100T on a single architecture.