Cisco has taken a strategic step in the race to dominate enterprise artificial intelligence infrastructure. The company announced in San Jose (California) the expansion of its Cisco Secure AI Factory with NVIDIA platform, now strengthened with the integration of VAST Data InsightEngine. The goal is clear: to enable a validated architecture capable of accelerating the use of agentic AI at scale, ensuring security, governance, and rapid data access.
This move aims to go beyond traditional chatbot approaches and open the door to AI agents capable of solving complex business problems in real time, based on up-to-date, protected corporate data.
The Promise of Enterprise Agentic AI
Agentic AI (or agentic AI) describes a model of AI systems capable of acting autonomously, communicating with each other, and collaborating with humans to achieve defined objectives. They need access to high-quality enterprise data, with low latency and strict security guarantees.
The challenge is that most current infrastructures are not prepared to support these workloads: bottlenecks in data flows, integration issues, and above all, risks of exposing sensitive information.
Cisco, together with NVIDIA and VAST, proposes a comprehensive design to address these limitations:
- Cisco AI PODs, building blocks based on UCS servers with GPUs NVIDIA RTX PRO 6000 Blackwell Server Edition.
- VAST InsightEngine, a key component of the VAST Data AI OS system, which converts raw data into AI-ready datasets.
- NVIDIA AI Data Platform, to accelerate computation and ensure low-latency interactions between models.
Validated Architecture: From RAG to Complex Reasoning
One of the strengths of the new architecture is the acceleration of RAG (Retrieval-Augmented Generation) workflows. These pipelines enable AI agents to retrieve information from databases or internal documents and combine it with generative models to provide accurate, up-to-date answers.
Cisco claims that its design reduces the latency of these processes from minutes to seconds, enabling responses almost in real time. This not only improves efficiency but also paves the way for agents capable of:
- Operating continuously and simultaneously across multiple workflows.
- Learning dynamically from interactions with data and users.
- Executing multi-step reasoning processes with contextualized results for business insights.
Additionally, the architecture incorporates role-based access controls, regulatory compliance, and integrated auditing, ensuring that innovation does not compromise security.
Strategic Partnerships: Cisco, NVIDIA, and VAST Data
The integration has been celebrated as a milestone by all three companies:
- Jeremy Foster, Senior Vice President and General Manager of Cisco Compute, emphasized: “The next generation of enterprise AI factories must be built on architectures that enable agents to access the right data at the right time. That’s the vision behind the Secure AI Factory.”
- Justin Boitano, Vice President of Enterprise AI at NVIDIA, noted that the next wave of agentic AI will depend heavily on enterprise data. “Connecting Cisco’s Secure AI Factory with NVIDIA and VAST InsightEngine provides the integrated platform companies need to deploy powerful AI agents at scale.”
- John Mao, Vice President of Strategic Alliances at VAST Data, highlighted that this collaboration represents “the first integrated design for large-scale RAG acceleration.”
Security in Every Token
One of the differentiators is Cisco’s concept of AI Defense, which applies security controls to every interaction and “token” generated by the models. This way, companies can trust that even in dynamic and distributed environments, sensitive information remains protected.
The integration with Splunk offers advanced visibility into the environments, with the ability to correlate events and reinforce compliance measures.
A Market in Full Bloom
Cisco’s initiative arrives at a pivotal moment: companies seek to leverage generative and agentic AI but face technical and regulatory challenges. Sector analysts suggest that differentiation will come from offering validated, scalable, and secure infrastructures capable of accelerating adoption without increasing risk.
The launch of AI PODs with VAST InsightEngine marks the first of a series of services Cisco plans to deploy for various enterprise AI use cases. This effort bolsters Cisco’s position against competitors that are increasingly talking about “AI factories” as the paradigm for the coming decade.
Availability
The Cisco AI PODs with VAST InsightEngine, validated according to NVIDIA’s enterprise AI platform reference design, are now available for order. The first service focuses on accelerating RAG pipelines, with more use cases expected to follow in the coming months.
Conclusion
The rollout of the Cisco Secure AI Factory with NVIDIA and VAST Data signifies a decisive step toward the industrialization of enterprise agentic AI. Combining computational power, advanced data management, and built-in security controls, Cisco provides organizations with a clear path to transform their data into intelligent, secure, and scalable agents.
According to the project leaders, it’s not just about building infrastructure but about shaping how companies will create the next generation of artificial intelligence.
Frequently Asked Questions (FAQ)
1. What is the Cisco Secure AI Factory with NVIDIA?
It’s a validated architecture that combines Cisco UCS servers with NVIDIA GPUs and VAST InsightEngine technology to enable secure, scalable deployment of AI agents in enterprise environments.
2. What role does VAST Data play in this solution?
VAST Data provides InsightEngine, which transforms raw data into AI-ready datasets, reducing latency in RAG pipelines and speeding up information availability for agents.
3. How does Cisco ensure the security of agentic AI?
Through AI Defense, which applies security policies and governance to every token generated by models, along with role-based access controls, regulatory compliance, and advanced visibility via Splunk.
4. What benefits does this offer compared to traditional solutions?
Drastic reduction in RAG process latency, support for multiple agents and workloads simultaneously, enterprise scalability, and security integrated directly into the infrastructure.
via: newsroom.cisco