Cisco is broadening its AI data center push, expanding its partnership with Nvidia and adding a new hardware ally in the process.
According to Cisco's own newsroom announcement, the company is extending its "Secure AI Factory with NVIDIA" for what it calls the rack-scale era — industry shorthand for the shift away from buying servers one box at a time and toward buying entire pre-integrated racks of compute, networking, and cooling as a single unit. Modern Nvidia AI systems draw enormous power and generate enormous heat, so they increasingly have to be designed, wired, and sold as a whole rack rather than assembled piecemeal.
Separately, a report carried on MSN describes Cisco partnering with Supermicro to add high-powered AI servers to its Nvidia-backed infrastructure offering. That detail matters because it marks Cisco expanding beyond networking — its traditional home turf — to sell more of the equipment needed to actually run AI workloads.
That is the strategic heart of the story. Cisco has long made its money on the switches and routers that move data between machines. Selling the servers that do the computing puts it closer to the center of AI infrastructure budgets, and closer to competitors that already sell full systems. Teaming with Supermicro lets Cisco offer those servers without building a comparable server line from scratch.
The sources here are limited to Cisco's announcement and the MSN report; neither excerpt provides pricing, availability dates, or deal terms, so those remain unknown.
Why it matters: as companies race to build AI capacity, the vendors who can deliver a complete, ready-to-run rack — not just one piece of it — stand to capture a far bigger share of the spending, and Cisco is now positioning itself as one of them.