NetApp and Cisco have expanded their collaboration around FlexPod with new validated architectures for AI workloads. The aim is to solve one of the most common tensions among organizations looking to move from proof of concept to full deployment: how to combine compute, networking, storage, security, and data management without building each platform from scratch.
Announced on June 3, 2026, during Cisco Live! in Las Vegas, this reinforces a long-standing partnership between the two vendors. FlexPod, the converged architecture traditionally based on Cisco and NetApp technologies, is now adapted to AI scenarios such as RAG, semantic search, inference, edge computing, and high-performance enterprise deployments with security controls integrated from the outset.
FlexPod adapts to production AI
Enterprise AI is entering a more demanding phase. Many organizations no longer settle for testing models in isolated environments or launching pilots with limited data. They want to bring use cases into production, connect models to corporate repositories, perform semantic searches, deploy internal assistants, accelerate inferences, all with security, data governance, and predictable performance.
This leap isn’t solely about choosing a model. Infrastructure must move data, power GPUs or XPUs, reduce bottlenecks, protect sensitive information, and enable repeatable operations. That is where NetApp and Cisco aim to position the new generation of FlexPod solutions for AI.
According to NetApp, the longstanding FlexPod partnership has already helped clients save up to 20% in management and maintenance of infrastructure. The company now presents this foundation as a starting point to simplify AI deployments, reduce risks, and avoid overly manual integrations.
| Scope | What the new FlexPod for AI offers | Why it matters |
|---|---|---|
| Enterprise AI | Validated architectures for RAG, semantic search, and scaled AI | Reduces complexity when moving from testing to production |
| Storage | NetApp AFX and disaggregated storage approach | Enables more independent scaling of performance and capacity |
| Data for AI | NetApp AI Data Engine, integrated with NVIDIA AI Data Platform design | Helps discover, prepare, and govern data for AI |
| Network | Cisco AI networking, Nexus One, and high-performance networking fabric | Aims to improve accelerator utilization and reduce execution times |
| Security | Cisco Secure AI Factory with NVIDIA and controls aligned with Zero Trust | Addresses risks related to data, compliance, and exposure |
| Inference and RAG | Pre-integrated solutions for departments and teams | Lower barriers in cost, complexity, and specialization |
| Edge computing | Cisco Unified Edge with NetApp storage options | Facilitates distributed AI with low latency and centralized management |
| Management | Proven architectures and automation | Enhances repeatability in complex deploys |
Data, network, and security: the less visible parts of AI
Enterprise AI is often explained through models, chips, or end-user applications, but the data layer significantly influences the outcomes. A company may have vast volumes of information, but without the ability to locate, classify, prepare, protect, and connect it to the right models, its value remains locked away.
NetApp introduces its AI Data Engine in this context—technology focused on discovering, preparing, and governing data within the infrastructure itself. It aligns with the NVIDIA AI Data Platform reference design, with the goal that data is ready to feed AI factories while maintaining control and traceability.
Cisco, on its part, provides the network and security layer. The reference to Cisco Secure AI Factory with NVIDIA points to an approach where security isn’t an afterthought but integrated into the architecture. In AI, this is particularly critical as risks extend beyond unauthorized access to encompass sensitive data exposure, governance lapses, improper model use, pipeline errors, leakage through queries, and compliance issues.
Networking also plays a crucial role. In training environments, intensive inference, or distributed processing, powerful accelerators alone aren’t enough. If the network introduces latency, performance drops, or unpredictable behavior, hardware investments can be underutilized. Cisco’s Nexus One aims to deliver a deterministic, high-performance network infrastructure to improve XPUs utilization, shorten job completion times, and increase result predictability.
The message is clear: production AI demands a more coordinated infrastructure. Storage, networking, security, compute, and data governance can no longer operate as separate pieces if the goal is to run critical workloads reliably.
Three scenarios: enterprise, departmental, and edge
NetApp and Cisco group their new solutions around three major use cases. The first is large-scale enterprise AI, suited for organizations seeking high-performance infrastructure for tasks like RAG and semantic search. This architecture enables AI capabilities close to where data resides—vital for companies that prefer not to move large volumes of sensitive information to external environments uncontrolled.
The second scenario focuses on inference and RAG flows for teams or departments. Not all companies need a full “AI factory” from day one. Many prefer to start with internal assistants, document analysis, knowledge retrieval, support automation, or tools allowing natural language queries of corporate data. Pre-integrated solutions can reduce the need for highly specialized staff and shorten the path to the first productive use case.
The third scenario targets edge computing. AI at the edge makes sense when data is generated far from the main data center or when latency is critical—industrial plants, retail stores, hospitals, logistics hubs, remote sites, or distributed infrastructure. Extending FlexPod to edge locations with Cisco Unified Edge and NetApp storage ensures a more cohesive operation instead of siloed technology islands at each site.
This is especially relevant as many organizations already run hybrid and distributed environments. AI won’t always be able to run solely in a data center or the public cloud. In some cases, deployment near data sources is necessary due to latency, cost, data volume, or regulatory requirements.
AI needs repeatable architectures
The movement by NetApp and Cisco reflects a broader market trend: organizations seek to reduce risk in AI projects through validated architectures. During early adoption, many set up proof-of-concept environments with disparate components. While useful for learning, this approach complicates scaling, security, auditing, and operational management in production.
FlexPod addresses this challenge with a proven approach: tested integration, enterprise-grade components, and repeatable deployment patterns. The innovation lies in tailoring this model for AI workloads, where performance, data, networking, and security demands are higher than in many traditional applications.
NVIDIA’s involvement adds another important layer. NetApp and Cisco highlight collaboration with NVIDIA to develop FlexPod solutions based on reference enterprise architectures. This supports the concept of “AI factories,” environments designed to build, deploy, and scale AI workloads with a controlled technical foundation.
World Wide Technology appears as a validation partner through its AI Proving Ground. For many companies, testing an architecture before full adoption is as crucial as vendor specifications. AI remains a promising but complex field, and customers need to validate performance, integration, and operational readiness before making significant investments.
The announcement doesn’t eliminate all common challenges in enterprise AI. Clear use cases, data preparation, team training, cost management, security policies, and ROI measurement remain essential. However, it indicates a maturing market: the conversation shifts from “which model” to “what infrastructure is needed to reliably implement AI.”
NetApp and Cisco aim for FlexPod to be the validated platform that accelerates AI projects without the burden of complex integration. As many organizations transition from lab experiments to production, this promise could translate into faster, safer, and more manageable deployments.
Frequently Asked Questions
What did NetApp and Cisco announce?
They announced new validated FlexPod solutions for AI workloads, focusing on performance, security, data management, and simplified enterprise deployment.
What AI use cases do the new FlexPod solutions address?
They target enterprise AI deployments, RAG, semantic search, inference, departmental workflows, and edge computing with virtualization, containers, and low-latency requirements.
What role does NVIDIA play in this initiative?
NetApp and Cisco collaborated with NVIDIA to build solutions based on reference enterprise architectures, including integration with AI Data Platform and Secure AI Factory concepts with NVIDIA.
Why is security important in AI infrastructure?
Because AI projects involve sensitive data, models, pipelines, and distributed components. Security must cover access control, data governance, compliance, and protection throughout the application lifecycle.
via: netapp