Cisco bulks up AI infrastructure portfolio with Supermicro liquid-cooled servers

“This expansion enables customers to easily manage complex, high-density AI clusters alongside non-AI workloads. Customers will also now be able to deploy rack-to-fabric liquid cooling, featuring Cisco liquid-cooled AI networking systems alongside Supermicro’s liquid-cooled servers. This unlocks trillion-parameter training and high-throughput inference use cases with platforms including Nvidia Vera Rubin NVL72 and Nvidia HGX Rubin NVL8,” Cisco stated. “By utilizing common operating models and existing technical competencies, organizations can implement a trusted AI infrastructure without the complexity of new, proprietary skill requirements.”

The expansion of liquid cooling options is an important component of the partnership. “Modern rack-scale systems, such as the Nvidia NVL72, can exceed 200 kW per rack, where liquid cooling becomes a system-level requirement rather than an option,” Cisco wrote in an FAQ about the announcement.

“Building on Cisco’s 100% liquid-cooled AI networking systems, liquid-cooled Cisco N9000 Series Switches interoperate directly with Supermicro’s rack-scale, liquid-cooled compute to deliver a rack-to-fabric liquid-cooled AI factory,” Cisco started. “This removes the thermal and power constraints that previously kept trillion-parameter training, high-throughput inference, and agentic AI out of reach for all but the hyperscalers. Also, it engineers the fabric for performance efficiency alongside the compute it serves.”