Cisco wants customers to ramp up AI adoption, so it’s partnering with Nvidia to bridge infrastructure hurdles

Cisco has partnered with Nvidia to provide integrated data center solutions in a move that a senior company executive said aims to help customers maximize the use of generative AI. 

The collaboration, announced at Cisco Live EMEA 2024, will combine Nvidia and Cisco’s hardware and software solutions, as well as professional services, to make it simpler for enterprises to harness the benefits AI can offer.

The partnership makes Nvidia’s latest Tensor Core GPUs available in Cisco’s M7 generation of UCS rack and blade servers, including the Cisco UCS X-Series, allowing better scalable performance for data-intensive AI use-cases.

Cisco’s validated reference architectures will make deploying and managing these AI clusters far easier, regardless of scale, the company said.

On stage at Cisco Live, Jonathan Davidson, EVP and general manager at Cisco Networking, emphasized how Cisco’s reference architectures were designed to save money in terms of energy and the efficiencies AI workloads produce.

“You all know Cisco designs its end to end reference architectures to take the guesswork out of designing and deploying automated data centers that run efficiently on workloads that help you save time and money.”

The partnership between the two firms will see Nvidia’s AI Enterprise software solution, with over 50 frameworks, pretrained models, and development tools, made available through Cisco’s global channel.

Nvidia and Cisco will also provide additional professional services through a set of trusted partners, which the company said will support enterprises lacking the knowledge or expertise required to deploy and manage their AI infrastructure securely.

Cisco wants to turn AI use cases into reality

This consideration is particularly important in supporting AI adoption among enterprises, according to Jeremy Foster, senior vice president and general manager of Compute at Cisco Networking.

Speaking to ITPro, Foster said the IT landscape is still relatively immature when it comes to AI and so providing the capabilities to actualize business use cases is an imperative.

“If you look at where we are with AI we are still very early in the adoption of these technologies for customers,” he said, citing Cisco’s 2023 AI Readiness Index that found 86% of organizations worldwide were not fully ready to integrate AI.

Foster explained businesses are rushing to adopt AI, and most can identify use cases for AI in their operations, but making this a reality requires a robust understanding of how the infrastructure for these technologies needs to be managed and deployed.

“Everyone sees the power available and has a use-case in mind, but how do you bring that to life? There’s a lot of things that happen between the concept and getting it out into a product,” Foster added.

This sort of expertise comes from experience, according to Foster, which businesses can access through Cisco’s partners.

“How do you train a model, fine-tune a model, and actually get it out to where it's delivering value to customers? The reason the services come in is because our key partners are doing this not one time for one business but 50 times for 50 businesses, and so they can really help reduce that time to value for customers.”