Dell Technologies Inc. today announced a slate of enhancements to its AI Data Platform with the aim of helping enterprises wrangle fragmented data and accelerate AI deployment.

The updates include deeper strategic partnerships with key players like Nvidia Corp., Elasticsearch BV and Starburst Data Inc.

At the heart of the announcement is Dell’s modular AI Data Platform, which decouples data storage from computing. The architecture is designed to smooth out performance bottlenecks and support demanding AI tasks such as model training, fine-tuning, retrieval-augmented generation and inferencing.

Enhancements to Dell PowerScale storage support Nvidia’s latest GB200 and GB300 graphics processing unit GPU NVL72 rack-scale server configurations. The PowerScale F710 model has also received Nvidia Cloud Partner certification for high-performance storage. Dell claims the F710 delivers GPU-scale performance with up to 72% lower power usage and five times greater space efficiency than competitors.

Dell’s ObjectScale storage platform now supports S3 object storage over using remote direct memory access, a technology that allows data transfers directly between the memory of two computers over a network bypassing the CPU and operating system. Dell said RDMA support, which enters tech preview in December, can deliver up to 230% higher throughput and 98% lower CPU usage. Coming enhancements to small object performance offer up to 19% higher throughput.

The announcement also brings new data engines designed to make AI data more usable. A new Data Search Engine, built with Elastic, allows users to query data using natural language for RAG and semantic search tasks. Integrated with Dell’s MetadataIQ metadata management software, it enables search across billions of files in both PowerScale and ObjectScale.

Dell’s Data Analytics Engine is gaining an Agentic Layer that uses large language models to automate data documentation and embed AI into SQL workflows. It supports multi-agent development and integrates with various vector databases, including PostgreSQL + PGVector and Dell’s own search engine.

Dell is also integrating Nvidia cuVS, an open-source library for GPU-accelerated vector search and data clustering, to bring GPU-accelerated hybrid keyword and vector search to its Data Search Engine.

The updated components will roll out between now and early 2026. PowerScale with Nvidia GB200 and GB300 support is available now, while ObjectScale S3 over RDMA and software updates will arrive in December. The Data Analytics Engine’s Agentic Layer and MCP Server will debut in February 2026 and both the Data Search Engine and Nvidia cuVS integration are expected in the first half of 2026.

Photo: Dell/X