Cisco and NVIDIA announced the intent to create a cross-portfolio unified architecture to simplify building AI-ready data center networks.
Cisco has unveiled an expanded collaboration with NVIDIA to deliver advanced AI technology solutions for enterprises. As businesses increasingly recognize AI’s critical role in driving growth, many are still in the early stages of adoption, facing challenges related to technical complexity and security in AI-ready data centers. This partnership aims to provide organizations with enhanced flexibility and options to meet the demands of AI workloads, ensuring high-performance, low-latency, and energy-efficient connectivity across data centers, cloud environments, and end-users.
The collaboration will integrate NVIDIA’s Spectrum-X Ethernet networking platform with Cisco’s and NVIDIA’s silicon technologies, forming the backbone for numerous enterprise AI applications. By enabling seamless interoperability between both companies’ networking architectures, the partnership focuses on delivering simplified, full-stack solutions tailored to customer needs. This initiative allows organizations to maximize their AI infrastructure investments through a unified architecture, utilizing existing management tools and processes across both front-end and back-end networks.
By facilitating the integration and standardization of Cisco and NVIDIA technologies, customers will gain access to current and future advancements within the NVIDIA Spectrum-X platform. These include features like adaptive routing, telemetry, congestion control, and low latency. Additionally, Cisco’s comprehensive networking, security, and digital resilience portfolio, encompassing the Splunk data platform, will further enhance these capabilities. As enterprises develop AI competencies and train models using their own data, implementing a robust security strategy becomes increasingly vital.
Also Read: Intel finalizes Xeon 6 lineup, covering all workloads.
In this expanded partnership, Cisco and NVIDIA are dedicated to introducing innovative solutions that accelerate AI adoption within the enterprise sector. Emphasizing the simplification of deployment and operations, as well as enhancing workload performance and visibility, the collaboration aims to deliver top-tier Ethernet solutions optimized for AI workloads. This includes extending NVIDIA’s Spectrum-X architecture to incorporate Cisco Silicon One, positioning Cisco as the exclusive partner silicon supported in Spectrum-X Ethernet solutions. Furthermore, the partnership seeks to transition enterprise data center Ethernet AI projects from prototype stages to full-scale deployments, which have traditionally been confined to public cloud or SaaS platforms.
Currently, Cisco and NVIDIA are planning to co-develop solutions for their clientele.
-
Development of Data Center Switches: Cisco intends to create data center switches utilizing the NVIDIA Spectrum Ethernet platform. This open ecosystem approach offers customers increased choice and flexibility, enabling organizations to standardize on the NVIDIA Spectrum-X networking platform with architectures based on both Cisco and NVIDIA switch silicon. This strategy combines industry-leading technologies from both companies under a unified management framework.
-
Collaboration on Reference Architectures: Cisco will work alongside NVIDIA to design and validate NVIDIA Cloud Partner (NCP) and Enterprise Reference Architectures. These will be based on NVIDIA Spectrum-X integrated with Cisco Silicon One, Hyperfabric, Nexus, UCS Compute, Optics, and other Cisco technologies.
Both companies are committed to ongoing collaboration and joint development efforts, aiming to provide customers with the highest-performing Ethernet solutions. Future initiatives will address critical challenges such as congestion management and load balancing, enabling customers to effectively scale and secure their AI deployments.
