What changed
Dell Technologies Inc. and AMD Inc. have announced enhancements to their enterprise AI infrastructure, including rack-scale systems that deliver 70% performance gains over previous generations. These systems can consolidate 13 servers from Dell’s 14th PowerEdge generation into a single unit. The collaboration also includes support for AMD Instinct MI350P PCI Express GPUs in Dell’s PowerEdge XE7745 and R7725 servers, offering 144 gigabytes of HBM3E memory.
What this means in practice
The new rack-scale systems are designed to meet the growing demand for general-purpose compute in agentic AI applications. By consolidating multiple servers into one, data centers can reduce physical space and power consumption while increasing computational density. Dell’s three-rack-unit server requires less power for cooling than a two-unit equivalent, allowing for more computing power per rack. The Dell AI Platform with AMD integrates AMD Instinct accelerators and EPYC processors with Dell PowerEdge servers, PowerSwitch networking, and PowerScale storage, enabling modular deployment and scaling for model training, fine-tuning, and high-throughput inference.
Limitations and availability
The Dell AI Platform with AMD is an on-premises solution that supports modular deployment, starting with smaller configurations and expanding over time. The platform includes AMD’s ROCm software stack for training and serving large language models. Support for AMD Instinct MI350P GPUs is available in Dell’s PowerEdge XE7745 and R7725 servers, starting in July. The collaboration also includes air-cooled GPU deployments, which are suitable for existing data center infrastructure. Dell and AMD have also partnered with the University of Cambridge and AT&T Inc. to advance AI research and telecom applications.
