Dell unveils new AI server family

Dell Technologies has unveiled a new range of AI-optimized servers powered by NVIDIA’s latest Blackwell Ultra chips, setting its sights on the surging enterprise demand for high-performance AI infrastructure.

Announced on Monday, the new Dell PowerEdge XE series servers come in both air-cooled and liquid-cooled versions and can support up to 192 NVIDIA Blackwell Ultra GPUs per system — and even scale to 256 GPUs per rack with Dell’s IR7000 infrastructure. According to Dell, these systems can train large language models (LLMs) up to four times faster than their predecessors.

“There’s a lot of interest in what’s next,” said Arthur Lewis, President of Dell’s Infrastructure Solutions Group, emphasizing the growing appetite for AI systems capable of handling ever-larger and more complex workloads.

As the AI arms race accelerates, Dell and competitors like Super Micro Computer are benefitting from a spike in demand for specialized AI hardware — though rising production costs and intense market competition continue to squeeze profit margins. Dell has already flagged a margin decline for fiscal 2026, while Super Micro recently cut revenue guidance, citing economic uncertainty and tariffs.

In response, Dell is doubling down on AI-centric storage, networking and managed services to secure profitability. The newly introduced servers are part of a broader strategy: to not just meet, but anticipate, enterprise AI infrastructure needs across industries like healthcare, manufacturing, and financial services.

What’s New in Dell’s AI Server Lineup:

• PowerEdge XE9780 and XE9785: Air- and liquid-cooled servers with up to 192 Blackwell Ultra GPUs.
• PowerEdge XE9712: Features NVIDIA GB300 NVL72 for 50x faster inference output and 5x throughput gains.
• PowerEdge XE7745: Coming in July 2025, will support NVIDIA RTX Pro 6000 Blackwell Server Edition GPUs, targeting robotics, digital twins and multimodal AI use cases.
• Future support planned for NVIDIA Vera CPUs and the Vera Rubin platform for scalable, next-gen AI compute.

On the networking front, Dell also added new switches to its AI infrastructure portfolio, including the Dell PowerSwitch SN5600 and SN2201 Ethernet switches and NVIDIA Quantum-X800 InfiniBand. These ultra-fast, low-latency switches deliver up to 800 Gbps throughput — essential for AI data center workloads.

Simplifying AI Operations

To support enterprise customers navigating AI’s complexity, Dell is rolling out new managed services for the Dell AI Factory with NVIDIA. These services cover everything from 24/7 infrastructure monitoring to AI platform management, version control and patching — addressing the talent shortage and operational bottlenecks many companies face.

Why It Matters

As businesses transition from experimental AI projects to full-scale deployment, the demand for enterprise-ready AI infrastructure is exploding. Dell’s partnership with NVIDIA positions it to be a major player in supplying the AI factories of tomorrow — the digital powerhouses driving decision-making and automation across sectors.