Dell unveils new AI hardware to boost performance

Dell Technologies has unveiled a series of new server, rack, and storage systems to bolster the Dell AI Factory, aimed at enhancing high-performance AI workloads.

Arthur Lewis, President, Infrastructure Solutions Group at Dell Technologies, emphasised the need for these advancements in data centres. "Today's data centres can't keep up with the demands of AI, requiring high density compute and liquid cooling innovations with modular, flexible, and efficient designs," he said. "These new systems deliver the performance needed for organisations to remain competitive in the fast-evolving AI landscape."

The Dell Integrated Rack 7000 (IR7000) has been introduced to address the demands of accelerated computing with advanced technologies. The rack supports high CPU and GPU densities with more sustainable power management and cooling techniques, aligning with Open Compute Project standards for large-scale applications. It has been designed to capture nearly 100% of the generated heat, supporting future deployments of up to 480KW cooling capabilities.

Designed with flexibility, the IR7000 integrates efficiently with both Dell and third-party networking solutions. It offers a fully integrated plug-and-play rack scale system optimised for AI workloads, simplifying deployment processes in various data centre environments.

For AI-ready platforms, Dell has introduced the PowerEdge XE9712, targeted for large-scale AI deployments and equipped with NVIDIA GB200 NVL72, allowing connections with up to 36 Grace CPUs and 72 Blackwell GPUs. This architecture supports real-time trillion-parameter large language model inferencing at a purported increase in efficiency of up to 25 times compared to traditional air-cooled systems.

Additionally, the Dell PowerEdge M7725 aims to deliver dense compute power, benefiting areas such as research, government, and higher education. Operable within the IR7000 rack structure, this system is built on 5th Gen AMD EPYC CPUs, offering scalable compute capabilities to meet the demands of intensive applications, ensuring energy efficiency via direct liquid cooling methodologies.

The enhancements to Dell PowerScale, positioned as the first Ethernet storage certified for NVIDIA DGX SuperPOD, aim to improve workload performance with advanced data management. Improvements focus on data discoverability through PowerScale metadata combined with the Dell Data Lakehouse, which supports operations by improving data ingestion time and decreasing associated costs.

The Dell Data Lakehouse updates aim to deliver more time-efficient operations with features such as disaster recovery and automated schema discovery. Customers can expect to reduce complexity in data management through enhanced Optimisation Services for Data Cataloging and Implementation Services for Data Pipelines, which prioritise structured integration and automation processes.

Aligned with the Dell AI Factory's initiatives, Dell Generative AI Solutions with Intel provide tailored platforms for AI deployment. It combines Dell server, storage, and networking resources with open-source software stacks, featuring Intel Gaudi 3 AI accelerators to deliver flexible AI solutions designed to cater to content creation and digital assistant use cases.

With these new developments, Dell Technologies continues its commitment to support digital transformation by strengthening the AI capabilities of enterprises globally, preparing them to handle emerging data challenges and opportunities.