Dell and HPE Expand AI Infrastructure with Nvidia-Powered Systems and Unstructured Data Solutions

Dell and HPE extend AI infrastructure lines with new Nvidia-powered systems - SiliconANGLE

Dell and HPE extend AI infrastructure lines with new Nvidia-powered systems Two of the largest enterprise infrastructure vendors announced product line expansions designed to give customers more options for deploying artificial intelligence workloads at scale. New offerings from Dell Technologies Inc. and Hewlett Packard Enterprise Co. include systems and software updates that incorporate Nvidia Corp.'s latest graphics processing unit technology and AI software, alongside tools from other partners, to help organizations move from AI experimentation to production. Dell said it's updating its AI Data Platform with new components aimed at improving unstructured data ingestion, transformation, retrieval and compute performance. The platform is designed to handle the full AI workload lifecycle, from data ingestion to agentic inference to knowledge retrieval. A new unstructured data engine developed in collaboration with Elasticsearch BV provides real-time, secure access to large-scale unstructured datasets, supporting advanced vector search, semantic retrieval and hybrid keyword search. The engine integrates GPU acceleration and Nvidia NeMo Retriever microservices for use in retrieval-augmented generation applications. The Dell AI Data Platform also integrates a federated SQL engine from Starburst Data Inc. for querying multiple data sources without moving the data, and an Apache Spark-based data processing engine for batch and real-time analytics. They sit alongside Dell's PowerScale and ObjectScale storage systems, which support structured, semi-structured and unstructured data. On the hardware side, Dell introduced PowerEdge R7725 and R770 servers equipped with Nvidia RTX PRO 6000 Blackwell Server Edition GPUs, which are data center-class graphics and compute processors built on Nvidia's Blackwell architecture and designed for AI inference, visual computing, engineering simulation and virtual workstation workloads. Dell said the new GPUs can deliver up to six times the token throughput for large language model inference compared to its previous generation, along with higher engineering simulation performance and support for more concurrent users. The R7725 will be the first 2U server to integrate the Nvidia AI Data Platform reference design, a validated blueprint for building GPU-accelerated infrastructure. Dell said the pairing yields a turnkey system for fast inferencing, semantic search and large AI workloads. HPE said it's expanding its line of AI-optimized systems with new configurations that also use Nvidia's Blackwell architecture GPUs, as well as the latest Nvidia AI models. The company said these updates are intended to support both agentic AI -- or systems that complete tasks autonomously or with little supervision tasks - and physical AI, which involves AI control of robotics and other hardware systems. HPE is also adding new options to its Cray XD670 and XD675 servers that allow them to be configured with Nvidia GB200 NVL72 systems for large-scale AI training, or with RTX PRO 6000 Blackwell GPUs for inference and smaller-scale workloads. HPE said both are pre-integrated with Nvidia's AI Enterprise software suite, including the NeMo framework and pretrained models such as Nvidia Nemotron for agentic AI and Cosmos for physical AI applications. The company is also updating its GreenLake for AI offering to support the new systems. GreenLake is a consumption-based model for deploying AI infrastructure, allowing organizations to scale GPU resources as needed while keeping workloads on-premises or in colocation facilities. HPE said its Private Cloud AI will support the latest versions of the Nemotron and Cosmos models, as well as the latest Nvidia Blueprint for Video Search and Summarization. Both companies' announcements center on integrating Nvidia's Blackwell-generation GPUs and related AI software into systems aimed at enterprises. Dell's focus is on connecting AI models to high-quality enterprise data, especially unstructured data, while HPE is emphasizing large-scale training and deployment of both agentic and physical AI models. The announcements are part of a broader trend by infrastructure vendors to provide turnkey AI infrastructure that combines hardware, storage, networking and AI frameworks in pre-tested configurations. Dell said its updated AI Data Platform with the unstructured data engine will be available later this year, alongside the PowerEdge R7725 and R770 servers with RTX PRO 6000 GPUs. HPE said its updated Cray XD systems and GreenLake AI configurations will also be available in the coming months.

Dell and Elastic Collaborate, Introduce 'Unstructured Data Engine'

It will offer advanced vector search, semantic retrieval and hybrid keyword search capabilities to gain relevant insights from data. Dell has announced updates to the Dell AI Data Platform, introducing a new 'unstructured data engine' that provides real-time access to large-scale unstructured data sets to gain relevant insights from. To realise these capabilities, Dell has entered into a collaboration with Elastic, a leading open-source AI search technology provider. This will offer users advanced vector search and semantic retrieval, and hybrid keyword search capabilities. This unstructured data engine operates in conjunction with other tools in Dell's AI Data Platform, such as a federated SQL engine for querying dispersed unstructured data, a processing engine for large-scale data transformations, and storage optimised for AI access. It will also leverage built-in GPU acceleration to achieve except

Dell, Elastic, Nvidia Partner For Unstructured Data Search, Retrieval

The partnership - announced Monday at SIGGRAPH 2025 in Vancouver - enhances the Dell AI Data Platform by automating data preparation to deliver natural language search capabilities inside data lakes that contain both structured and unstructured files. Dell Technologies has partnered with Elasticsearch and Nvidia to build server and storage arrays capable of tackling data search and processing within the vast hoards of information which has become a bottle-neck to AI adoption. The partnership - announced Monday at SIGGRAPH 2025 in Vancouver - enhances Dell's AI Data Platform by automating data preparation to deliver natural language search capabilities inside data lakes that contain both structured and unstructured files. "The key to unlocking AI's full potential lies in breaking down silos and simplifying access to enterprise data," said Arthur Lewis, president of Dell's Infrastructure Solutions Group, in a statement. "Collaborating with industry leaders like Nvidia and Elastic to advance the Dell AI Data Platform will help organizations accelerate innovation and scale AI with confidence." [RELATED: Dell's AI Charge: Think Faster, Act Smarter ] The new capabilities are built on the Nvidia AI Data Platform reference design and the Elastic product Elasticsearch. When new data is ingested into the system, these engines immediately index every asset, ensuring it is accessible through natural language search, integrated with dashboards for end users, and processed for quick turnarounds. It is made possible through a new collaboration with open-source search AI leader Elastic, Dell said. It offers customers advanced vector search, semantic data retrieval -- in which the search engine understands the intention behind the queried information -- and hybrid keyword search capabilities and takes advantage of GPU acceleration to boost performance. "Fast, accurate, and context-aware access to unstructured data is key to scaling enterprise AI," Ken Exner, chief product officer at Elastic, said in a statement provided by Dell. "With Elasticsearch vector database at the heart of the Dell AI Data Platform's unstructured data engine, Elastic will bring vector search and hybrid retrieval to a turnkey architecture, enabling natural language search, real-time inferencing, and intelligent asset discovery across massive datasets." The technology is built on three software engines, Dell said. The first, a federated SQL engine, is for instant queries across multiple data sources -- such as data lakes, databases, and object storage -- using standard SQL and delivering unified analysis without moving data. The second is the unstructured data engine which is powered by Elasticsearch. This gives users fast full-text, semantic and vector searches so they can instantly access documents, or video clips across massive archives. Third is the data processing engine that uses Apache Spark for accurate, real-time analytics around ETL -- the extract, transform, load processes used to prepare data for analytics - and the machine learning processes that build models around that data. To power those engines, Dell has released validated designs for its air-cooled 2U PowerEdge R7725 and R770 servers with Nvidia RTX Pro 6000 Blackwell GPUs and PowerScale and ObjectScale storage arrays. The PowerEdge R7725 is the first 2U server to integrate the Nvidia AI Data Platform reference design. Once it is paired with the Dell AI Data Platform and its new unstructured data engine, it becomes a turnkey solution for the enterprise, Dell said. "Enterprises worldwide need infrastructure that handles the growing scale and complexity of AI workloads," said Justin Boitano, vice president of enterprise AI at Nvidia, in a statement. "With Nvidia RTX PRO 6000 GPUs in new 2U Dell PowerEdge servers, organizations now have a power efficient, accelerated computing platform to power AI applications and storage on NVIDIA Blackwell."

Dell Updates AI Data Platform with NVIDIA and Elastic to Simplify Enterprise AI Workflows

Dell Technologies announced updates to the Dell AI Data Platform to help customers better support the full lifecycle of AI workloads from ingestion and transformation to agentic inferencing to AI-powered knowledge retrieval. Enterprise data is massive, growing rapidly and increasingly unstructured, but only a fraction of it is usable for generative AI today. To unlock its value, organizations need continuous indexing and a vector retrieval engine that converts content into embeddings for fast, precise semantic search. As workloads grow, organizations need infrastructure that streamlines data preparation, unifies data access across silos and delivers end-to-end enterprise-grade performance.