Photo by panumas nikhomkhai on Pexels

Dell Technologies is ramping up its AI capabilities with a new generation of servers leveraging Nvidia’s Blackwell Ultra GPUs. These AI acceleration platforms promise a significant boost in performance, with Dell claiming up to four times faster AI training speeds compared to previous offerings.

The new PowerEdge XE9780 and XE9785 servers, available in both air-cooled and liquid-cooled (XE9780L and XE9785L) versions, are designed for diverse deployment environments, from traditional data centers to high-density, whole-rack setups. These systems can house up to 192 Nvidia Blackwell Ultra GPUs with direct-to-chip liquid cooling, scalable to 256 GPUs per Dell IR7000 rack.

“Our goal is to make AI more accessible,” said Michael Dell, chairman and CEO, highlighting the company’s ambition to empower enterprises to manage the end-to-end AI lifecycle through the Dell AI Factory with Nvidia.

Beyond the servers themselves, Dell’s strategy incorporates a comprehensive ecosystem encompassing networking, storage, and software. The PowerSwitch SN5600 and SN2201 switches, alongside Nvidia Quantum-X800 InfiniBand switches, form the networking backbone. The Dell AI Data Platform has been updated for optimized data management. Software offerings include the Nvidia AI Enterprise platform, featuring Nvidia NIM, NeMo microservices, and Blueprints. Dell is also introducing Managed Services for the AI Factory with Nvidia, offering monitoring, reporting, and ongoing maintenance.

While the rollout is planned throughout 2025, industry analysts note that Dell’s success hinges on its ability to deliver demonstrable business value to clients. The close partnership with Nvidia also presents a potential risk, creating a dependence on Nvidia’s component supply chain.