Dell Technologies has expanded the Dell AIデータプラットフォーム with context orchestration layers, GPU-accelerated processing engines, and high-density multitenancy across its unstructured storage infrastructure. The architectural enhancements target the enterprise retrieval path, connecting raw files, databases, and multi-site repositories directly to AI agent pipelines. The system integrates NVIDIA ネモトロン Retriever models for document parsing, embedding, and reranking, while NVIDIA cuVS provides accelerated vector indexing and search routines. Context components run locally within customer data centers, maintaining data governance without external model or storage vendor lock-in.
Context Orchestration and Retrieval Engines
The architecture introduces three context-mapping components to resolve taxonomy mismatches across disparate internal systems. The Unified Semantic Layer provides consistent definitions through centralized rules, searchable glossaries, and support for imported industry ontologies or enterprise taxonomies. Dell is also enabling NVIDIA Auto-Ontology, an open-source library that builds knowledge graphs from enterprise data, to extend the semantic layer. Working in tandem, the Enterprise Knowledge Graph continuously maps lineage, metadata, and historical query behavior across both structured tables and unstructured documents.
When an agent executes a query, the platform limits traversal to authorized tables, multimodal data products, and vector stores, letting agents correlate events such as anomalous telemetry with specific equipment logs, supply batches, and customer orders. Knowledge Agents operate as single-topic advisory instances bounded by defined segments of the Knowledge Graph, running under user-enforced compute spend caps, security policies, and retrieval accuracy thresholds, with Nemotron Retriever models supplying their reasoning and visual understanding.
“Data without context is just noise. Most enterprises have spent years making their data accessible. That’s not the same as making it usable,” said Arthur Lewis, president of the Infrastructure Solutions Group at Dell Technologies. “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.”
Data Path GPU Acceleration and PowerScale Multitenancy
Dell integrates NVIDIA acceleration stacks directly into its データ処理エンジン to minimize data preparation latency. Per internal vendor testing, the Dell Data Processing Engine running NVIDIA cuDF on NVIDIA RTX PRO 4500 ブラックウェル サーバー エディション GPUs processes data nearly 4 times faster on average than CPUs alone across a mixed workload suite, achieving up to 20 times faster processing on batch processing workloads.
Dell’s footnote puts the figures at a 3.9x average and a 20.4x peak on a data mining workload, measured on GPU-accelerated versus CPU-only Apache Spark runs on a PowerEdge R770 with default configurations and no tuning. Apache Arrow lets jobs query data in place between Dell storage targets and local compute, which Dell says can cut data-preparation time.
“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,” said Jason Hardy, vice president of storage technology at NVIDIA.
On the storage tier, Dell PowerScale clusters now scale to 500 isolated tenants per cluster, backed by mutual TLS (mTLS) over NFS for data-in-transit encryption and authentication and granular role-based access control per tenant, aimed at AI service providers and enterprises running shared AI platforms. Infrastructure teams can also use the open-source Dell Storage Performance Tool to benchmark S3-compatible object storage throughput and latency under AI training, checkpointing, and inference workloads. Dell is expanding AI Data Platform implementation services across its data and storage engines to cover analytics, processing, search, and orchestration tuning.
Release Cadence and Availability Status
The Dell Storage Performance Tool and Dell AI Data Platform implementation services are available now. PowerScale cluster multitenancy and mTLS over NFS enhancements are slated for release in November 2026. The Dell Data Processing Engine enhancements with the NVIDIA acceleration stack will arrive in December 2026, followed by Apache Arrow acceleration integration in 1H 2027. The Unified Semantic Layer, Enterprise Knowledge Graph, and Knowledge Agent context frameworks are scheduled for release in 1H 2027.
Amazon