Avatar Avatar

Organizations around the world are looking to modernize and scale their data centers to account for surging growth in artificial intelligence (AI). But right now, 86% of companies are not fully prepared to leverage AI and AI-powered technologies to their fullest potential, according to the Cisco AI Readiness Index.

The fact is, AI is driving a rethinking and retooling of data center infrastructure and management. It requires enterprises to operate at hyperscale to process increasingly large data sets for training and inferencing AI tools. Customers also need to understand that infrastructure requirements vary in terms of building, optimizing, and using AI models—each with different performance demands across workloads.

Overcoming these challenges demands a new way to operationalize, simplify, and scale data center infrastructures to meet the evolving demands of AI, as well as other business-critical compute and networking-intensive workloads. Organizations understand there are sizeable benefits in productivity, efficiency, security, and intelligence using generative and predictive AI tools—not to mention a boost to day-to-day operations and overall business value.

That’s where Cisco Networking Cloud can help. With new and enhanced solutions and technologies announced at Cisco Live, we are empowering companies to simplify their infrastructure operations at scale for full-stack AI implementations.

Deploy and manage AI infrastructure at scale—with ease

Customers are looking to simplify deployment and operations across the full stack of AI infrastructure through integrated data, compute, software, storage, and high-performance networking. Cisco, in collaboration with NVIDIA, has developed the Cisco Nexus HyperFabric AI cluster solution, which is designed to simplify deployments for the enterprise with plug-and-play cloud management for AI-native infrastructure. With Cisco Nexus HyperFabric AI clusters, customers will be able to tap the full stack of NVIDIA MGX and VAST Data storage, in addition to the high-performance Cisco Ethernet infrastructure.

The Cisco Nexus HyperFabric AI cluster solution guides organizations through the entire infrastructure lifecycle process, from design to order, deploy, configure, validate, and continuous operations for both AI and non-AI deployments. This gives a broad set of IT and non-IT personnel, including DevOps and application teams, the ability to effortlessly deploy and manage every aspect of the Ethernet fabric and AI cluster management with a single click.

For those looking to simplify configuration, monitoring, and maintenance of all tenant customer fabrics—whether they are Cisco ACI, Cisco NX-OS, or an AI/ML fabric—we have announced Cisco Nexus one fabric experience. This solution makes private cloud/on-premises management easier than ever to support a variety of data center network use cases across multiple fabric technologies using the Cisco Nexus Dashboard as a single control point. In addition, end-to-end segmentation further helps unify fabrics and simplify their management through common policies, which eliminates data silos and helps reduce attack surfaces.

To make it easier to manage complex workloads like AI that traditionally warrant multiple applications, we have consolidated the different Nexus Dashboard applications into common services, which will significantly streamline software install and service upgrades while requiring less compute resources for hosting.

These Nexus Dashboard enhancements, serving as our private cloud/on-premises operations and automation platform for data center networks, introduce capabilities such as topological views for greater network visualization, plug-and-play management capabilities for faster deployment, switch-level energy management, AI-enabled root cause analysis for improved assurance, and AI/ML workload visibility to pinpoint and quickly remediate AI performance issues.

To help companies more easily deploy, scale, and upgrade hyperconverged clusters with a sustainable, future-ready solution for intensive workloads like AI, Cisco Compute Hyperconverged with Nutanix on UCS X-Series is the industry’s first hyperconverged solution on a modular blade architecture. Built on UCS X-Series, Cisco’s most popular blade computing system with 5th Gen Intel® Xeon® Processors inside and one hundred percent cloud-managed by Cisco Intersight, this solution with Nutanix delivers unparalleled flexibility, simplicity, and resiliency to customers looking to simplify their hybrid, multicloud environments in an AI-driven world.

We’re also committed to giving customers more choice of high-performance servers with compelling advantages for AI workloads based on their unique use case, performance, and sustainability needs.

We’re adding the UCS X215c modular blade built with 4th Gen AMD EPYC™ Processors to our lineup of M8 UCS servers managed by Cisco Intersight. This system can support up to 2,048 cores in a fully populated seven RU UCS X-Series system, making it among the densest computing platforms in the market today—and making Cisco the only top-tier server vendor in the industry to offer a blade system with AMD. This launch comes on the heels of two recently announced UCS M8 rack servers built with AMD, including our UCS C245 M8 server, which already set 22 world records for performance, making it a versatile platform powerful enough to run any application with ease. Customers who upgrade to UCS M8 servers with AMD can see up to 2.8x performance over the previous generation.

Easily leverage AI-native tools with validated designs

Deploying AI-ready infrastructure can be complex, time-consuming, and costly. The Cisco Validated Design (CVD) for GPT-in-a-Box on Cisco Compute Hyperconverged with Nutanix is a turnkey, AI-ready solution designed to help companies simplify the deployment process, jumpstart AI initiatives, and accelerate time to value.

Because the CVD has already been tested and validated with the most popular large language models (LLMs), customers face less deployment risk and can expedite AI-project delivery. And with prescriptive steps to deploy a full-stack AI platform, IT teams don’t require AI-specific expertise. As a complete AI solution, GPT-in-a-Box enables organizations to deploy their hyperconverged solutions in an array of environments—from centralized data centers to remote edge sites.

Another way we are making it easier for companies to experience fast, reliable, and fully predictable AI/ML deployment is with Nexus 9000 CVD for CPU/DPU-based AI clusters. This CVD provides high interoperability and predictability of a GenAI infrastructure that achieves high-performance, low latency, and lossless GenAI fabrics with Cisco Nexus 9000 Series Switches, Cisco Optical Networking, and the Intel Gaudi 2 AI accelerator—managed by Cisco Nexus Dashboard.

You can find all of our CVDs for AI-ready infrastructure in the Cisco Validated Design Zone.

Reap the value of AI

These new and enhanced Cisco Networking Cloud capabilities enable organizations to modernize, scale, and simplify their data center infrastructures to seamlessly take AI/ML projects from the drawing board to production scale. Companies can implement and manage full-stack AI with revolutionary cloud-based operations and private cloud/on-premises managed solutions that guide customers through the lifecycle—from value creation to value realization across all their AI clusters.

See how you can prepare your data center for the future.


Kevin Wollenweber

SVP/GM, Cisco Networking

Data Center and Provider Connectivity

Jeremy Foster

Senior Vice President & General Manager

Cisco Compute