Cisco, Nvidia Partner On Joint Architecture For The Creation Of AI Data Centers

Building on their February announcement, the two companies have revealed Cisco Secure AI Factory with Nvidia with two different deployment options for enterprises.

Technology artificial intelligence digital ai hand concept on cyber future business tech science innovation futuristic network strategy background virtual data communication learning assistant search.

Cisco Systems and Nvidia are building on their existing relationship with the introduction of an “AI factory” architecture with security at the heart of the offering, the two companies revealed on Tuesday.

Cisco Secure AI Factory with Nvidia will help enterprises build and secure data centers to develop and run AI workloads, “dramatically” simplifying the process for these businesses, the two companies said.

The announcement builds on Cisco and Nvidia’s February partnership news that involved the joint development of Ethernet-based networking offerings for AI data centers.

The new offering will “transform” AI factories meticulously designed to power AI workloads, using co-optimized architecture from Cisco and Nvidia, Cisco’s Chief Product Officer Jeetu Patel said in a blog post published on Tuesday on the news.

“Our partnership with Nvidia is expected to offer organizations unmatched flexibility and agility, embedding security from the application and the workload to the underlying infrastructure. Whether you’re at the beginning of your AI journey or looking to scale existing infrastructure, Cisco Secure AI Factory with Nvidia will provide scalable, high-performance solutions that adapt to your unique requirements. With customizable modular components and proven reference architectures, we intend to empower you to design and deploy AI infrastructure that aligns seamlessly with your business goals,” Patel said in the post.

There will be two deployment options available, including the vertically integrated deployment option and the modular deployment option. The vertically integrated deployment option is based on Cisco Nexus Hyperfabric AI and includes Cisco 6000 Series Switches, accelerated compute with Cisco UCS C885A M8, based on the Nvidia HGX platform with Nvidia H200 GPUs, Nvidia BlueField-3 DPUs and SuperNICs, the Nvidia AI Enterprise software platform, VAST Data Storage, and Cisco Optics.

The modular deployment option, on the other hand, will include customizable components, such as Cisco UCS C885A M8 and C845A M8 servers, based on the Nvidia HGX and MGX platforms with Nvidia H200 GPUs, and managed by Cisco Intersight, Nvidia BlueField-3 DPUs and SuperNICs, Cisco Nexus 9000 Series switches managed by the Cisco Nexus Dashboard, future Cisco switches with Nvidia Spectrum-X, managed by Nexus Dashboard, the Nvidia AI Enterprise software platform, storage from Nvidia-Certified partners Pure Storage, Hitachi Vantara, NetApp, and VAST Data, Open Source containerization and automation solutions from Red Hat, and Cisco Optics.

“AI factories are transforming every industry, and security must be built into every layer to protect data, applications and infrastructure,” said Jensen Huang, Nvidia’s founder and CEO said in a statement. “Together, Nvidia and Cisco are creating the blueprint for secure AI—giving enterprises the foundation they need to confidently scale AI while safeguarding their most valuable assets.”

All deployment options will include protection via Cisco’s security offering that are powered by the company’s 2024 purchase of Isovalent, a leader in open source, cloud-native networking and security, including Cisco Secure Firewall, Cisco Hypershield and Cisco AI Defense, the company said.

Offerings based on the Cisco Secure AI Factory with Nvidia architecture are expected to be available for purchase before the end of calendar year 2025, the companies said.