Cisco Expands Nvidia-Based Secure AI Factory with Supermicro in 2026

Cisco announced on Tuesday a partnership with Supermicro to expand its Nvidia-based Secure AI Factory solution, adding high-density liquid-cooled and air-cooled computing systems. Under the plan, Cisco will integrate Supermicro's rack-scale, high-density GPU servers into its AI infrastructure portfolio, packaging them with its networking, security, and observability products as pre-validated systems to help customers deploy and manage large-scale GPU clusters while accommodating non-AI workloads.

The collaboration also covers liquid-cooling solutions from rack to switching fabric. Cisco will combine its liquid-cooled AI networking equipment with Supermicro's liquid-cooled servers to support scenarios such as trillion-parameter model training and high-throughput inference. Cisco's president and chief product officer, Jeetu Patel, stated that Cisco is positioned as 'critical infrastructure for the AI era' and will support various architectural choices. Patel emphasized that whether customers use frontier models from hyperscale cloud providers, open-weight models, or choose on-premises deployment or edge-based AI, Cisco can provide the corresponding infrastructure.

In terms of market performance, Cisco has largely avoided the 'SaaS doom' selloff that hit software stocks in 2026, with its shares rising nearly 45% since the start of 2026. This is partly attributed to the company's focus on hardware. The expansion also brings Cisco further into a more competitive market, where its partners and some customers compete in certain areas. Patel downplayed this tension, noting that the overlapping business scope is relatively limited and the complementary potential is greater.