Cisco just made its biggest bet yet on becoming the plumbing of the AI era. The company announced a partnership with Supermicro on August 25 to integrate NVIDIA-based AI compute systems into its Secure AI Factory, creating what amounts to a one-stop shop for enterprises looking to deploy AI infrastructure without stitching together five different vendors’ hardware.
The pre-validated systems, which combine Supermicro’s high-density servers with Cisco’s networking, security, and observability stack, are expected to hit availability in October 2026.
What Cisco is actually building
The Secure AI Factory first debuted in March 2025 at NVIDIA’s GTC event in San Jose, positioning Cisco’s networking and security expertise as the connective tissue for AI workloads. By March 2026, the platform had expanded to include advanced edge computing capabilities and multi-agent AI functionalities.
Now, with Supermicro in the picture, Cisco is adding the raw muscle. Supermicro brings liquid-cooled and air-cooled server designs purpose-built for NVIDIA GPUs. Rather than asking customers to figure out how Supermicro boxes talk to Cisco switches, the partnership delivers pre-validated configurations where everything is tested to work together out of the gate.
