Cisco unveiled a new data center switch optimized for AI workloads that was developed with help from Nvidia.

The Cisco N9100 switch was introduced at Nvidia's GTC event in Washington, D.C. It features Nvidia’s Spectrum-X ASICS, leveraging its integrated remote direct memory access (RDMA) and advanced packet congestion controls to boost AI workload performance.

The switch uses Cisco's Silicon One and Cloud-scale offerings, including the recently revealed P200 chip, and is targeted at the rapidly expanding AI data center market, with Cisco having one eye on marketing it for the nascent neocloud space.

“We're at the beginning of the largest data center build-out in history,” Cisco Chief Product Officer Jeetu Patel proclaimed. “The infrastructure that will power the agentic AI applications and innovation of the future requires new architectures designed to overcome today's constraints in power, computing, and network performance.”

Customers can deploy the switch with either Cisco's NX-OS or the open-source SONiC operating system, providing operators building AI infrastructure with greater flexibility and choice.

“Spectrum-X Ethernet delivers the performance of accelerated networking for Ethernet,” Gilad Shainer, SVP of networking at Nvidia, noted. “Working with Cisco’s cloud reference architectures and Nvidia cloud partner design principles, customers can choose to deploy Spectrum-X Ethernet using the newest Cisco N9100 series or Cisco Silicon One-based switches to build open, high-performance AI networks.”

The related reference architecture will also include other networking technologies unveiled at GTC this week, including the fourth-generation BlueField data processing unit (DPU) and ConnectX-9 SuperNIC.

The N9100 series switch is scheduled for availability by year-end.