Cisco and Nvidia Expand Infrastructure Partnership to Deliver Rack-Scale AI Systems to Enterprises

Cisco Systems Inc. has expanded its Cisco Secure AI Factory collaboration with Nvidia Corp. to supply Nvidia-certified rack-scale system architectures directly to enterprise data centers and regional cloud operators, according to analysis first reported by SiliconANGLE. The initiative aims to lower operational complexity for organizations building high-density artificial intelligence environments outside of major public cloud networks.

The rollout comes as spending on artificial intelligence hardware remains heavily concentrated among a small group of technology conglomerates. According to figures from CreditSights, five major U.S. hyperscalers—Amazon Web Services Inc., Microsoft Corp., Google LLC, Meta Platforms Inc., and Oracle Corp.—are projected to direct between $700 billion and $800 billion toward total capital investments this year, with approximately three-quarters of those funds targeted directly at AI infrastructure projects.

A separate study from Allianz determined that hyperscale operators and large tech service providers currently absorb roughly two-thirds of all artificial intelligence expenditures worldwide. To support the training of frontier artificial intelligence models, these tech giants have constructed bespoke, liquid-cooled computing facilities, creating compute supply constraints for conventional corporate technology divisions.

However, infrastructure demand is beginning to broaden beyond public cloud platforms toward distributed AI infrastructure builders. This emerging segment encompasses specialized neocloud operators, sovereign cloud initiatives, dedicated AI service providers, and standard enterprise data centers seeking to run local workloads.

Enterprise technology executives face substantial technical and strategic barriers when building modern compute environments. Executing complex tasks, including trillion-parameter artificial intelligence models and multi-agent software frameworks, requires advanced liquid cooling and dense rack configurations. At the same time, corporate data governance regulations and competitive considerations prevent many organizations from transmitting sensitive intellectual property to public cloud models, while an ongoing shortage of specialized networking engineers complicates custom graphics processing unit deployments.

The expanded Cisco and Nvidia partnership addresses these operational obstacles by introducing standardized, Nvidia-certified rack-scale architecture blueprints. Rather than requiring internal engineering teams to manually assemble complex hardware configurations, the systems offer pre-integrated liquid cooling and enterprise management mechanisms designed to simplify deployment timelines.

The standardized design also provides an operational framework connecting local corporate infrastructure with external cloud providers. As neoclouds and sovereign cloud operators deploy regional capacity built on Cisco systems, enterprise IT teams can temporarily burst processing tasks from on-premises Cisco AI Factory hardware to external Cisco-powered clouds while maintaining consistent security protocols, network settings, and management policies.

The joint offerings mark a transition in the AI hardware market from customized hyperscale installations toward repeatable enterprise data center standards, offering corporate IT organizations a path to operate high-density compute hardware with simplified governance and operational controls.