Cisco Adds Rack-Scale Computing to AI Factory Portfolio
Networking giant partners with Supermicro to deliver liquid-cooled, high-density GPU systems for neocloud and sovereign cloud deployments.
Cisco is extending its AI infrastructure portfolio beyond networking into rack-scale computing through a partnership with Supermicro, addressing surging demand from neocloud and sovereign cloud providers building out massive AI capacity.
The expansion adds high-density, liquid- and air-cooled server systems to Cisco's Secure AI Factory architecture, which already includes the company's networking switches and management software. Customers will be able to deploy rack-to-fabric liquid cooling that pairs Cisco's liquid-cooled networking gear with Supermicro's cooled servers, enabling trillion-parameter model training on platforms including NVIDIA Vera Rubin NVL72 and NVIDIA HGX Rubin NVL8.
Cisco first reported the partnership details on August 25, 2026, positioning the move as a response to one of the largest datacenter buildouts in history. The company will begin offering Supermicro compute as part of its AI Factory portfolio in October 2026.
Why it matters
The partnership marks a strategic shift for Cisco, traditionally a networking vendor, into the full-stack AI infrastructure market dominated by server manufacturers and cloud providers. By bundling compute, networking, and liquid cooling into validated reference architectures, Cisco is betting that customers—especially emerging neocloud operators and government-backed sovereign clouds—will pay a premium for integrated solutions that reduce deployment complexity and timeline risk. The move also positions Cisco as the only NVIDIA technology partner using its own networking switches in NVIDIA Cloud Partner compliant solutions.
Full-stack architecture for AI workloads
The expanded offering combines Cisco Silicon One-based switches for front-end networking with NVIDIA Spectrum-X switches for back-end AI cluster connectivity, unified under Cisco's Nexus One management platform. The architecture supports NVIDIA Cloud Partner compliance requirements, a key consideration for neocloud providers building multi-tenant AI infrastructure.
Cisco emphasized three customer benefits: reduced supply chain risk through combined expertise from Cisco and Supermicro; faster deployment through new Cisco Validated Infrastructure Services that certify installations match reference architectures; and simplified operations through integration with NVIDIA AI Enterprise software and Cisco Cloud Control for end-to-end observability.
The company is establishing a dedicated large-scale AI lab to develop certification tools and test software for infrastructure validation. This service offering aligns with NVIDIA's own Infrastructure Services program.
Targeting neocloud growth
Jeetu Patel, Cisco's president and chief product officer, framed the expansion around customer demands for data control and return on investment alongside raw performance. The neocloud and sovereign cloud segments represent particularly strong growth opportunities, as governments and enterprises seek alternatives to hyperscale public clouds for sensitive AI workloads.
Justin Boitano, NVIDIA's vice president of enterprise AI, characterized AI infrastructure as "revenue-generating" rather than cost centers, with compute producing intelligence that drives business value.
The partnership leverages Supermicro's manufacturing scale and liquid cooling expertise while allowing Cisco to maintain its networking differentiation. For customers, the integrated approach promises to simplify procurement and reduce the technical risk of assembling multi-vendor AI clusters from scratch.
Details were first reported by Cisco in a company press release.
This is an original analysis by the Omega editorial team. Source reporting: AI Watch.
Want systems like this working for your business?
Book a Call