The Cisco Secure AI Factory with NVIDIA brings together the infrastructure, security, and observability enterprises need to move AI into production. At the core of that architecture are Cisco AI PODs, the modular, pre-validated building blocks that help customers deploy and scale enterprise AI with greater speed, consistency, and confidence.
That is why Cisco’s new validated design for the VAST AI OS on Cisco AI PODs matters. This CVD validates the Cisco EBox running VAST AI OS as the shared data foundation for Cisco AI PODs. It is the foundational first step toward certifying the world’s first unified AI Data Platform, beginning with the layer every enterprise AI deployment depends on first: persistent, high-performance, shared data infrastructure.
This momentum did not start here. In 2025, Cisco introduced VAST InsightEngine on Cisco AI PODs as part of its expanding Cisco Secure AI Factory with NVIDIA strategy, giving customers a validated path to accelerate real-time retrieval, RAG pipelines, and agentic AI workflows on Cisco infrastructure. That solution was built on Cisco UCS C845A M8 servers, powered by AMD EPYC processors, and optimized for NVIDIA RTX PRO 6000 Server Edition GPUs. That launch established the AI-native data intelligence layer. This new CVD builds on that progress by validating the persistent shared data foundation with Cisco EBox. The next step is to bring that foundation together with Cisco CNode-X, extending the architecture from shared data infrastructure into a more complete unified AI Data Platform designed to bring compute to data.
This first step is important because AI factories need more than accelerated compute. They need a data foundation that can keep GPUs fed, simplify operations, and make enterprise data continuously available for AI. In the Cisco Secure AI Factory with NVIDIA, Cisco AI PODs provide the modular infrastructure base, NVIDIA provides the accelerated compute and AI software foundation, and VAST provides the unified data layer that turns enterprise data into an always-available asset for retrieval, reasoning, and agentic workflows.
This CVD starts at exactly the right layer. The initial focus is validating the Cisco EBox cluster for use as the persistent shared data foundation on Cisco AI PODs. Built on Cisco UCS, Cisco EBox provides the durable all-flash data plane for VAST AI OS. That gives customers a validated starting point for consolidating enterprise data on one high-performance foundation before layering on richer AI data services.
That shared foundation matters because fragmented AI environments create hidden costs. When storage, metadata, retrieval services, analytics engines, and inference pipelines are spread across separate systems, data has to be copied, moved, transformed, and re-indexed across multiple layers. That introduces latency, duplication, operational drag, and governance complexity. VAST’s approach is built to break those trade-offs by reducing unnecessary data movement and bringing data services closer to the data itself on a shared architecture.
This is also where the relationship between Cisco EBox and Cisco CNode-X becomes strategically important.
Cisco EBox establishes the persistent shared data foundation. Cisco CNode-X brings GPU-accelerated data services directly to that foundation. Together, they are engineered to bring compute to data, eliminating the inefficiency of pushing data through fragmented layers for vector search, analytics, and AI pipelines. That architectural model is what enables a truly unified AI Data Platform.
In practical terms, that means the architecture is designed to evolve in a deliberate sequence. The Cisco EBox validates the shared persistent foundation first. The VAST InsightEngine on Cisco UCS C845A M8 was the first ever VAST Certified CNode-X platform in the market, and demonstrated how AI-native retrieval and data intelligence can run on Cisco AI PODs using AMD EPYC and RTX PRO 6000 Server Edition acceleration. Together, they point toward the next phase of the architecture, where storage, vector operations, analytics, event-driven processing, and AI data services operate on one integrated foundation instead of being stitched together from disconnected products.
That broader direction is already visible in VAST’s platform approach. InsightEngine delivers real-time vector embedding and retrieval for RAG and agentic workflows. DataEngine extends the platform with event-driven execution and orchestration that moves compute close to data using event triggers and built-in streaming capabilities. The result is a model that unifies files, objects, tables, vectors, streams, and metadata in one architecture rather than forcing enterprises to assemble separate infrastructure tiers for each function.
That is why this CVD should be viewed as more than a storage validation. It validates the first layer of a larger engineering design. Today, that means certifying Cisco EBox with VAST AI OS as the shared data foundation for Cisco AI PODs. Next, that foundation expands with Cisco CNode-X and the proven VAST InsightEngine on Cisco UCS C845A M8 approach to bring vector search, accelerated analytics, event-driven services, and richer AI data pipelines directly into the platform. That is the path toward the broader Cisco AI Data Platform CVD expected in the coming months.
This also aligns cleanly with the broader Cisco operating model around AI infrastructure. Cisco AI PODs are already positioned as foundational building blocks of Secure AI Factory with NVIDIA, and Cisco has connected that architecture to Intersight-based operations, Splunk Observability Cloud, and a broader secure cloud-native framework. The result is not just a high-performance infrastructure stack, but a modular path toward a more unified and more operationally complete enterprise AI platform.
For customers, the message is straightforward. This CVD establishes the validated shared data foundation for Cisco Secure AI Factory with NVIDIA today. It gives enterprises a production-aligned way to start with the data layer first, simplify infrastructure, and reduce the hidden tax of fragmented data movement. And it points directly to what comes next: a broader unified AI Data Platform built on the combination of Cisco EBox and Cisco CNode-X, VAST InsightEngine accelerated by the Cisco UCS C845A M8 showing how AI-native retrieval services fit into that architecture and VAST AI OS bringing persistent data and accelerated data services together on Cisco AI PODs.
To learn more about Cisco’s approach to AI infrastructure validation, visit the Cisco Validated website and look under Design guides for AI-ready infrastructure.
If you’re at GTC, visit the Cisco (#1421) and VAST (#1007) booths to see how VAST AI OS on Cisco AI PODs helps power the data foundation for Cisco Secure AI Factory with NVIDIA, and get a preview of how Cisco EBox and Cisco CNode-X come together in the next step toward a unified AI Data Platform.
