Dell will have an eight-way Nvidia Blackwell server with liquid cooling — coming later this year

Nvidia
(Image credit: Nvidia)

When Nvidia announced its Blackwell B200-series processors earlier this spring, the company also introduced its own servers based on the GPUs. While Nvidia confirmed that traditional server makers like Dell, HPE, and Lenovo will also offer Blackwell-based machines, there weren't many details at the time. That changes with today's announcement from Dell.

Nvidia and Dell took the wraps off the upcoming Dell PowerEdge XE9680L server with liquid cooling and eight Nvidia Blackwell Tensor Core GPUs. Liquid cooling implies that the PowerEdge XE9680L server will be aimed at applications that require the highest performance possible, so we expect the machine to use the B200 GPU and cost a lot of money.

It's noteworthy that Dell already has a PowerEdge XE9680 machine based on Intel's 4th Generation Xeon Scalable CPUs, with support for Nvidia's H100 processors and AMD's Instinct MI300X accelerators. The L version will likely use the same basic design, only it will be capable of supporting Nvidia's Blackwell GPUs for AI and HPC workloads, which probably means upgrading the system with higher power PSUs and better cooling.

Interestingly, Dell and Nvidia call the Dell PowerEdge XE9680L server 'one of industry's densest, energy-efficient rack-scale solutions for large Blackwell GPU deployments.' Hence, we are indeed talking about an ultra-powerful solution for large-scale AI deployments.

Speaking of large-scale AI deployments, the so-called AI factories are certainly among them. The Dell AI Factory, introduced in partnership with Nvidia earlier this year, delivers comprehensive AI solutions spanning from data centers to edge environments, facilitating rapid and large-scale AI adoption for organizations. This platform combines Dell hardware expertise with Nvidia's software technologies, offering customers a broad AI portfolio and an open ecosystem of technology partners.

"We are unleashing this super genius power. Everyone is going to have access to this technology — and it is going to get smarter," said Michael Dell, chief executive of Dell.

By leveraging this collaboration with Nvidia, the Dell AI Factory aims to establish AI sovereignty for companies and countries, ensuring robust data security and enabling the development of tailored AI services.

"Our partnership between us is really about that, literally from the ground up building AI factories and delivering it to the world’s enterprises as a solution," said Jensen Huang, chief executive of Nvidia.

Anton Shilov
Contributing Writer

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

  • Interestingly, Dell and Nvidia call the Dell PowerEdge XE9680L server 'one of industry's densest, energy-efficient rack-scale solutions for large Blackwell GPU deployments.'

    Yeah. Not surprising though, since it comes in the compact 4U form factor, industry’s highest density x86 server solutions. Most likely candidate is obviously next-gen NVIDIA HGX B200.

    Pretty sure the 'L' means liquid cooling here, since XE9680L leverages direct liquid cooling/DLC smart cooling for both CPUs and GPUs. Btw, DELL claims customers can scale up to 72 Blackwell GPUs per 52 RU rack, or 64 GPUs per 48 RU rack.

    Also, it appears that while XE9680 has 10 PCIe Gen 5 FHHL slots, this "L" variant has 12 PCIe Gen 5 FHHL slots, so that means 2x capacity for high-speed I/O for North/South AI fabric as well (if we consider a config of 8 PCIe Gen 5 FHHL slots 1:1 allocated to the 8 GPUs in the XE9680 and XE9680L).
    Reply