NVIDIA and Cisco have teamed up to develop Ultra-Dense Small Form Factor Pluggable (USFP), a new optical transceiver form factor designed to support 400 Gb/s-per-lane connectivity in next-generation AI factories. The technology promises four times the bandwidth density of current 200 Gb/s-per-lane pluggable optics, with improved power efficiency and capacity of up to 12.8 Tb/s. The development is significant for European data center operators, network equipment makers and optical component suppliers looking to scale AI infrastructure without increasing network complexity or power consumption. It also opens up opportunities for connectivity manufacturers as the industry moves toward higher-density Ethernet networks. Four times the bandwidth density USFP is designed to help AI infrastructure operators connect more computing systems within the same network footprint, avoiding the additional switching layers that can increase latency, costs and power consumption. The first planned transceivers will integrate 16 full-duplex lanes into an OSFP-sized footprint, supporting both switches and SuperNICs. Future versions could expand to 32 lanes or more. According to the companies, USFP offers four times the bandwidth and front-panel bandwidth density of existing 200Gb/s-per-lane pluggable modules, alongside approximately 1.3 times better power efficiency. “Building the world’s largest AI factories starts with rethinking the smallest of connections,” said Gilad Shainer, senior vice president NVIDIA. “With USFP, NVIDIA and Cisco are building an ecosystem to reinvent pluggable optics — connecting more computing with greater efficiency to power the next generation of AI factories.” New connector design targets power and reliability A key innovation is a two-dimensional array connector featuring a two-step insertion mechanism. High-speed electrical contacts engage only after the module is inserted, helping reduce contact wear, improve signal integrity and provide more predictable insertion force. The design supports electrical signaling at 400 Gb/s per lane while also improving performance at 200 Gb/s. USFP modules can dissipate up to 100W without requiring a dedicated embedded cooling circuit, potentially simplifying switch designs and improving reliability. The architecture also supports optical technologies that simplify signal processing to reduce latency and power consumption, while accommodating different connection distances. Building an open optics ecosystem NVIDIA and Cisco are working with TE Connectivity and Amphenol to develop USFP connectors, cages and modules, with more partners expected to join. The companies also plan to establish an open industry specification through a community-led multi-source project, giving component suppliers a common platform for developing interoperable high-density optical hardware. The initiative builds on existing collaboration between NVIDIA Spectrum-X Ethernet networking and Cisco Silicon One-based networks, targeting AI infrastructure deployments across enterprises, neocloud providers and sovereign cloud operators. For European AI infrastructure developers, the proposed standard could provide a more flexible route to higher-bandwidth networking while preserving the serviceability and deployment advantages of pluggable optics. The post NVIDIA and Cisco unveil USFP optics for AI factories appeared first on eeNews Europe .