NVIDIA and Cisco Plan 400-Gb/s-per-Lane USFP Optics for AI Networks
The proposed USFP form factor targets 12.8 Tb/s modules and denser Ethernet fabrics; its open industry specification remains planned.
NVIDIA and Cisco announced plans on October 12 for Ultra-Dense Small Form Factor Pluggable, or USFP, an optical module form factor designed for 400-Gb/s-per-lane electrical signals in AI data-center networks.
The first planned transceivers would combine 16 full-duplex lanes in a footprint comparable to an octal small form-factor pluggable module, supporting up to 12.8 Tb/s overall capacity. The companies describe scope to increase lane counts in later designs.
NVIDIA reports four times the front-panel bandwidth density and about 1.3 times the power efficiency of 200-Gb/s-per-lane pluggable modules. Those are supplier design claims, not independently measured results reported by Infraday. The aim is to connect more computing systems without adding network tiers and their associated power and complexity.
TE Connectivity and Amphenol are collaborating on cages, connectors and modules. NVIDIA and Cisco plan an open industry specification through a future community-led multi-source project. The announcement does not establish a completed specification or a general-availability shipping date; operators should distinguish the proposed ecosystem from deployed production equipment.
NVIDIA and Cisco technical announcement — October 12, 2026 · AI Infrastructure. Official media contact: press@nvidia.com. Source reviewed October 15, 2026.