Cisco and NVIDIA have jointly unveiled the Ultra-Dense Small Form-Factor Pluggable (USFP), the industry’s first pluggable form factor designed to scale AI networks. This innovation aims to provide up to 12.8T of aggregate capacity while maintaining existing deployment and serviceability models. As AI workloads become more complex, the demand for bandwidth, responsiveness, and scalability has surged. The USFP addresses these needs by offering a path to support both 200G and 400G SerDes-based systems and ASICs, ensuring seamless integration with current infrastructure.
The USFP is engineered to handle the growing traffic across data centers, clouds, and edge environments while balancing power, cooling, and operational flexibility. With 16 lanes capable of operating at 400 Gb/s per lane, it delivers 6.4T of aggregate capacity, four times the bandwidth density of today’s 1.6T pluggables. The architecture can scale to 32 lanes, potentially reaching 12.8T of aggregate capacity, and offers approximately 1.3x greater power efficiency through 400G SerDes technology.
Cisco and NVIDIA emphasize that the future of AI networking will not rely on a single architectural approach. The USFP extends the benefits of independently field-replaceable modules, mature supply chains, and interoperability, ensuring operational simplicity and ecosystem flexibility. This collaboration reflects a shared goal of accelerating AI adoption while maintaining openness and choice for customers.
The USFP is part of Cisco’s broader platform vision, combining silicon, high-density systems, optics, software, and operational expertise. This approach ensures that customers can scale AI infrastructure efficiently and sustainably without redesigning their existing setups. Cisco and NVIDIA remain committed to delivering innovations that advance performance while preserving trusted operational foundations.