Dell Integrates NVIDIA GB300 “Blackwell Ultra” GPUs into AI Factories for Enhanced Performance and Scalability

In the wake of NVIDIA’s recent keynote at Computex, Dell has unveiled a range of advanced AI solutions powered by NVIDIA’s cutting-edge AI architecture. These innovations aim to elevate the “AI factory”ecosystem, reinforcing Dell’s commitment to being a leader in AI infrastructure.

Expanding Dell’s AI Portfolio with NVIDIA Technologies

Over recent years, Dell has significantly enhanced its AI infrastructure, especially following the introduction of NVIDIA’s Hopper generation solutions, which have seen widespread adoption across various AI markets. This momentum is set to continue, as Dell is proud to announce its latest advancements with new PowerEdge AI racks.

These new racks are equipped with NVIDIA’s Blackwell GPUs, specifically the HGX B300, enabling configurations that can support up to 192 GPUs. This capacity ranks them among the most powerful solutions currently available. Moreover, Dell is distinguished as one of the first companies to integrate NVIDIA’s RTX PRO 6000 “Blackwell”GPUs into its offerings.

Dell’s new servers include the PowerEdge XE9780 and XE9785 models, which are designed with air cooling and can be outfitted with up to 192 Blackwell Ultra HGX B300 GPUs. In addition, the liquid-cooled variants, the PowerEdge XE9780L and XE9785L, can accommodate an impressive 256 GPUs per rack, optimizing performance for high-demand environments.

The air-cooled versions, Dell PowerEdge XE9780 and XE9785, facilitate integration within existing enterprise data centers, while the liquid-cooled PowerEdge XE9780L and XE9785L accelerate deployment at rack scale. Featuring compatibility with up to 192 NVIDIA Blackwell Ultra GPUs and customizable options for 256 GPUs per Dell IR7000 rack, these solutions outperform previous models, delivering up to four times faster training for large language models (LLMs).

The Dell PowerEdge XE9712, equipped with NVIDIA GB300 NVL72, is engineered for efficiency in AI training, boasting a 50-fold increase in AI reasoning output and a 5x improvement in throughput, thanks to the new Dell PowerCool technology.

Scheduled for release in July 2025, the Dell PowerEdge XE7745 will utilize NVIDIA RTX Pro™ 6000 Blackwell Server Edition GPUs. This platform is being positioned to accommodate a range of AI applications, including robotics and digital twins.

Dell is also working on supporting the NVIDIA Vera CPU, promising enhancements in speed and efficiency.

Additionally, the company plans a new Dell PowerEdge XE server tailored for NVIDIA’s Vera Rubin platform, designed for scalable integrated rack systems.

This strategic focus on innovation has led Dell to achieve an astounding fourfold increase in the speed of LLM training compared to previous generations, underscoring its position as a leading integrator in the landscape of AI technology.

Additionally, Dell is set to introduce an AI solution powered by NVIDIA’s latest RTX PRO Blackwell GPUs. The anticipated PowerEdge XE7745, targeting the industrial sector particularly for multi-modal AI applications, is due for release by July 2025. The firm has broader plans that include developing PowerEdge designs utilizing NVIDIA’s Vera CPUs and Rubin GPUs.

“We’re on a mission to bring AI to millions of customers worldwide. Our job is to make AI more accessible. With the Dell AI Factory and NVIDIA, enterprises can manage the entire AI lifecycle across use cases, from deployment to training, at any scale.”

– Michael Dell, CEO of Dell Technologies

With a robust NVIDIA-focused AI portfolio and groundbreaking Blackwell Ultra products, Dell is well-positioned to push the boundaries of what’s possible in the AI space, enhancing its “AI factory” initiatives.

Source & Images