Dell expands AI platform with semantic layer for agents

Dell has expanded its AI Data Platform with new data orchestration, processing and storage features for AI agents. The update adds a semantic layer and knowledge graph intended to give applications a consistent view of enterprise data.

The changes address a common problem in corporate AI projects: useful information is often scattered across files, databases, cloud services and internal systems that were not structured for machine use. Dell aims to connect those sources inside a customer's own data centre so AI agents can draw on governed internal data rather than rebuild context from scratch for every query.

At the centre of the update is a Unified Semantic Layer, designed to give structured and unstructured information consistent business meaning through rules, definitions and a searchable glossary. That should help AI systems treat different labels for the same concept as equivalent, such as recognising that one system's "client" may be another system's "account".

The platform also adds an Enterprise Knowledge Graph that maps relationships across structured and unstructured data. Dell says it uses metadata, lineage and query history to refine the graph over time, allowing applications and agents to locate related tables, data products, multimodal data and vector indexes based on a user's access permissions.

A third addition, called Knowledge Agents, is intended to apply that context to specific topics. According to Dell, each agent can be limited to a defined section of the knowledge graph, with customers setting rules for guidance, data access, quality thresholds and spending limits.

Dell says these three elements will remain inside the customer environment because they contain sensitive and valuable information. The platform is designed to share context across applications and agents without tying users to a single model, data source or storage provider.

Processing gains

Dell also outlined changes to the data path behind AI workloads. Its Data Processing Engine, using NVIDIA cuDF on NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, can process data nearly four times faster on average than CPU-only configurations across a range of workloads, with gains of up to 20 times on batch processing tasks.

Apache Arrow is being used to move data between Dell storage and processing systems so jobs can query data in place. Dell says this can reduce data preparation time before information is used by AI systems.

On the storage side, Dell is extending PowerScale to support up to 500 tenants in a single cluster. The update also adds mTLS over NFS for encrypted and authenticated file traffic, along with more granular role-based access controls for tenant management in shared AI environments.

These changes target service providers and large organisations running shared AI platforms for multiple teams or customers on the same infrastructure. Dell says the model is intended to let them isolate data between tenants while using a single cluster.

Benchmarking tool

Another part of the announcement is the Dell Storage Performance Tool, which is available now. It is designed to help customers size AI infrastructure, compare supplier systems and test S3-compatible object storage performance across training, inference and checkpointing workloads.

Dell is also expanding its implementation services around the AI Data Platform. The services cover the platform's analytics, processing, search and orchestration components, with the aim of helping organisations move from deployment to production use.

Arthur Lewis, President, Infrastructure Solutions Group, Dell Technologies, said: "Data without context is just noise. Most enterprises have spent years making their data accessible. That's not the same as making it usable. An agent that can find a customer record but doesn't know what it means, how it connects to everything else, or whether it can be trusted isn't intelligent. It's just fast. The companies that solve it won't just deploy AI faster or run more agents. They'll get more out of the data they already have."

Dell says the Storage Performance Tool and AI-ready data services are available now. PowerScale multitenancy and security updates are due in November 2026, Data Processing Engine enhancements with the NVIDIA acceleration stack are scheduled for December 2026, and the Unified Semantic Layer, Enterprise Knowledge Graph and Knowledge Agents are set for release in the first half of 2027.

Jason Hardy, Vice President of Storage Technology at NVIDIA, said: "AI agents are only as effective as the data they can access, understand, and trust, and NVIDIA accelerated computing and AI software help turn governed data into AI-ready context. By bringing NVIDIA cuDF acceleration directly into the Dell Data Processing Engine, Dell helps shorten the path from stored data to GPU-accelerated AI-ready data their agents can use to drive real world innovation."