Scaling on-premises AI with Dell and AMD infrastructure

Dell is expanding its enterprise AI offering through integration of AMD GPU technology with its PowerEdge servers. Starting in July 2026, the PowerEdge XE7745 and R7725 models will support AMD Instinct MI350P PCIe GPUs, aiming to enable increased AI performance without requiring data centre redesign.

In the current business environment, enterprise AI is becoming increasingly important. On-premises solutions are used to provide performance, control, and security, and are playing a growing role in enterprise infrastructure decisions. Dell and AMD are contributing to this shift with jointly enabled infrastructure options for enterprise use.

The Dell AI Platform, enhanced with AMD technology, is designed to support the scaling of AI workloads from pilot stages through to production. With support for AMD Instinct MI350P GPUs, PowerEdge servers provide an option for organisations looking to scale generative and agentic AI workloads within existing environments.

With AMD MI350 PCIe GPUs, Dell PowerEdge servers provide:

• Seamless deployment: Standard air-cooled PowerEdge XE7745 and R7725 servers, with no data centre redesign required
• Leading edge performance: Up to 4,600 peak teraflops (MXFP4) and 144GB HBM3e memory
• Diverse workload capability: Optimised for small, medium, and large model inference, RAG pipelines, and agentic AI applications
• Open software integration: AMD enterprise AI software supports frameworks including PyTorch, TensorFlow, and vLLM, with minimal code changes required and no licensing fees

The Dell and AMD collaboration has also resulted in a modular platform designed to support evolving enterprise AI requirements. Organisations can configure systems based on current workloads and scale compute and GPU density over time without requiring full system redesign.

The platform incorporates AMD Enterprise AI Suite, AMD ROCm, and AMD Inference Server to support AI lifecycle processes including training, fine-tuning, inference, and agentic workflows within a secure, validated on-premises environment.

For higher intensity workloads, Dell also offers PowerEdge XE9785 servers with AMD MI355X GPUs and EPYC CPUs, designed for foundation model development and large-scale inference, extending the range of enterprise AI deployment options.