Cisco, NVIDIA, and VAST Data Advance Agentic AI Infrastructure with Secure AI Factory Blueprint

Cisco has unveiled a new blueprint for enterprise-scale AI infrastructure, designed to unlock agentic AI workloads with security, speed, and governance at the core. The announcement expands the Cisco Secure AI Factory with NVIDIA to include new use cases, most notably the acceleration of retrieval-augmented generation (RAG) pipelines. This enables faster data extraction and retrieval, allowing AI agents to securely access the information they need in real-time.

Building Blocks for Agentic AI

The foundation of this architecture is Cisco AI PODs, which now come tightly integrated with VAST InsightEngine, a key component of the VAST Data AI OS. These AI PODs transform raw enterprise data into AI-ready datasets, leveraging the NVIDIA AI Data Platform reference design to streamline performance and deployment.

Inside these PODs, Cisco’s UCS server portfolio, powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, delivers high-performance compute tailored for next-generation AI applications. Cisco RTX PRO Servers are among the first systems to implement NVIDIA’s AI Data Platform reference design, strengthening Cisco’s position as an enterprise-ready provider of AI infrastructure.

On the networking side, Cisco’s high-performance Ethernet ensures seamless connection across compute and data nodes. Combined with NVIDIA accelerated computing and AI software, this unified stack reduces latency, enabling AI models and agents to interact more responsively with business data.

Jeremy Foster, SVP and GM of Cisco Compute, emphasized the transformational potential of agentic AI, noting that Cisco’s work with NVIDIA and VAST was central to building an architecture that supports the next chapter of enterprise AI adoption.

Justin Boitano, VP of Enterprise AI at NVIDIA, added that the future of agentic AI depends on securely tapping into enterprise data for grounding and context during inference. He highlighted the collaboration as a way to create a unified platform for scalable AI agent deployment.

John Mao, President of Strategic Alliances at VAST Data, called the integration of VAST InsightEngine into Cisco’s AI PODs “a pivotal advancement.” Mao added that this collaboration marks the beginning of a new era of secure, collaborative AI agents operating at an unprecedented scale, particularly with the first integrated design optimized for RAG acceleration.

Addressing Enterprise AI Infrastructure Demands

Agentic AI introduces unique infrastructure challenges. Enterprises aim to deploy AI agents that interact with humans and with other agents to solve complex, multi-step problems. To enable this, infrastructure must deliver:

• Reduced latency in RAG pipelines, cutting data retrieval times from minutes to seconds for near-real-time AI responses.
• Scalability by design, so multiple agents and workloads can run concurrently with high throughput—empowering advanced reasoning and dynamic learning at enterprise scale.
• Security and governance are baked into the architecture, including role-based access controls, compliance features, and audit readiness, ensuring that sensitive data is protected while still fueling AI innovation.

By combining compute, networking, storage, and governance into a unified solution, Cisco’s Secure AI Factory provides enterprises with a trusted framework for deploying agentic AI at scale.

Availability

Cisco AI PODs with VAST InsightEngine, powered by the NVIDIA AI Data Platform, are now available to order from Cisco. The RAG-optimized design is the first in a broader series of AI Service PODs designed to support diverse enterprise use cases across various industries.