The Synergy of Cisco and Nvidia: A New Era of AI Infrastructure

In a groundbreaking collaboration, Cisco and Nvidia have entered into a strategic partnership to revolutionize the world of artificial intelligence (AI) infrastructure. Their joint efforts aim to provide customers with integrated software and networking solutions that will streamline the process of building AI-supportive infrastructures.

One of the core objectives of this collaboration is to emphasize the significance of Ethernet networks in managing AI workloads within enterprises. By aligning their strategies, Cisco and Nvidia are strengthening their position in the development of Ethernet networks for AI applications. This partnership also grants both companies access to their respective sales and support systems, ensuring seamless customer integration.

Cisco will leverage its extensive global distribution network to sell Cisco Ethernet-based solutions specifically designed for Nvidia. This comprehensive portfolio will include professional services and support provided by key partners, facilitating the implementation of GPU clusters through Ethernet infrastructure.

To further enhance their offerings, Cisco will incorporate the latest Nvidia Tensor Core GPUs into their Cisco Unified Computing System (UCS) rack servers and blade systems. These cutting-edge products are tailored to handle AI workloads and process large volumes of data in data centers and at the network edge. The integrated package, set to be released in the second quarter, will include Nvidia AI Enterprise software, featuring pre-trained models and developer tools for ready-to-use AI solutions.

To ensure streamlined deployment and management of AI clusters, Cisco has introduced Cisco Validated Design (CVD)-validated reference architectures. These reference architectures provide a comprehensive blueprint for deploying AI clusters at any scale, suiting a wide range of applications. This includes virtualized and containerized environments, incorporating converged and hyperconverged options. Cisco plans to release CVDs for FlexPod and FlashStack for AI inferencing with Nvidia AI Enterprise software later this month, with more architecture options to follow.

The integrated package will encompass deep integration with other Cisco software. For example, ThousandEyes Digital Experience Monitoring software will be integrated to provide AI-driven insights and automatic issue remediation between the cloud and on-premises networks. This seamless integration will enhance the overall performance and reliability of AI infrastructure.