Dell Technologies and silicon partner AMD introduced the Dell PowerEdge XE9680 Server equipped with the AMD Instinct MI300X Accelerator and ROCm 6 open software platform. This server is designed for heavy machine learning, deep learning, and generative AI workloads, promising to accelerate tasks in data centers and support AI model training.
The collaboration between AMD and Dell, as discussed by AMD’s Director of HPC/AI Strategic Accounts Justin Ionescu and Dell’s Director of Strategic Global Partnerships Suresh Raam, aims to enhance the cloudification of telco networks during their transition to 5G Standalone (SA) and the decentralization of compute from data centers to the network edge.
“AMD and Dell are very much focused on democratizing the AI space for both hardware and software,” remarked Ionescu. Raam further noted that this democratization should begin at the telco data centers. He emphasized that virtualization starts at these centers, making it an ideal location to promote network operational benefits and modernization.
From the telco perspective, building on-premise data centers offers several advantages. Security, latency, and data sovereignty are better managed than with a public cloud infrastructure. Raam cited studies suggesting that telcos could reduce costs by up to three times by opting for on-prem data centers, factoring in operational costs, leasing infrastructure, and software licenses.
Moreover, Dell has identified certain AI applications within telco data centers that necessitate millisecond latencies. These real-time operational applications can’t achieve such performance with cloud-based architectures. Therefore, a key goal of the Dell and AMD partnership is to build the idea that on-premise data centers are the future.
Ionescu views AI as a pervasive technology mega-trend extending from endpoint to edge to cloud. He notes that it propels early adopters into advantageous positions.
