Dell has been busy on the AI infrastructure front in recent months, detailing its work with Nvidia and teaming up with upstart and Nvidia challenger Cerebras. Most recently, AMD was the beneficiary of Dell’s AI attention, as Dell announced shipping availability of its Dell PowerEdge XE9680 AI servers with AMD Instinct MI300X accelerators.
At the same time, Dell published its Dell Validated Design for Generative AI with AMD, a design guide which was announced in May, and is aimed at supporting enterprise implementations of the PowerEdge with AMD accelerators for AI inference, fine-tuning and retrieval-augmented generation. Dell also said it will have more deployment support services coming in September to help customers to help smooth integration of the Dell-AMD system into enterprise environments.
The PowerEdge-MI300X offering features eight MI300X accelerators, a combined 1.5 TB of HBM3 memory and 42 petaFLOPS of peak theoretical FP8 with sparsity precision performance, according to Dell.
The shipping announcement from Dell is another momentum-builder for AMD’s AI ambitions and more specifically for the MI300X, which became available late last year and is also playing a key role in a Microsoft AI offering, according to an announcement from Microsoft Build in May.
Like many large companies helping enterprises build out their AI infrastructure, Dell is not just betting on one chip partner to help it get the job done. The company said in mid-May that it was launching Nvidia-supported AI Factory offerings, including a liquid-cooled version of the PowerEdge XE9680 that will use eight of the Nvidia Blackwell GPUs announced last spring. Dell founder Michael Dell also said late last month that Dell is working with Nvidia to build an AI Factory to support Elon Musk's xAI Grok chatbot.
In another recent AI move by Dell, the company announced that it is partnering with GPU upstart Cerebras on memory storage systems for AI that also include PowerEdge R6615 servers and AMD EPYC CPUs.