HPE unveils ProLiant Gen13 servers for AI workloads
Thu, 8th Oct 2026 (Today)
HPE has introduced four ProLiant Gen13 servers powered by 6th Gen AMD EPYC processors, the first wave of its next server line for AI and other compute-intensive enterprise workloads.
The new systems are the ProLiant DL525 Gen13, DL585a Gen13, XD245 and XD285. They are designed to expand compute, memory and input-output capacity while adding new security and management features for distributed IT environments.
Two of the servers target mainstream enterprise deployments that need more processing density for inference, analytics and virtualised workloads. The other two are built on Open Compute Project standards for denser rack-scale deployments handling data-intensive AI and high-performance computing workloads.
According to HPE, the DL585a Gen13 is a high-density platform that supports up to eight double-wide PCIe GPUs, two 6th Gen AMD EPYC processors and PCIe Gen6 connectivity. It is intended for demanding uses including AI inference, retrieval-augmented generation model fine-tuning, analytics, virtualisation and agentic AI.
The DL525 Gen13 is a 1U single-socket system with one 6th Gen AMD EPYC processor and up to 256 cores. It supports 16-channel memory, including DDR5 and MRDIMM, as well as PCIe Gen6, and is positioned for high-throughput tasks such as fraud detection, market data processing, risk modelling, simulation and AI inference.
Rack density
The XD245 and XD285 are aimed at larger-scale compute environments. According to HPE, the XD245 uses a one-open-rack-unit, four-node direct liquid-cooled design with support for up to eight 6th Gen AMD EPYC processors, while the XD285 uses a two-open-rack-unit, two-node air-cooled design for a wider range of data centre environments.
Those systems target CPU-intensive workloads including computer-aided engineering, weather and climate modelling, scientific research and financial modelling. Their design focuses on fitting more compute into a smaller footprint while improving power distribution and cooling.
The launch comes as hardware suppliers seek to position conventional server platforms for a broader range of AI uses beyond model training. HPE linked the new systems to rising enterprise demand for infrastructure that can run inference and support more autonomous software systems while still handling established workloads such as databases, analytics and virtualisation.
"Organisations need to turn their investments in agentic AI and inference into business value without adding operational complexity or risk," said Krista Satterthwaite, Senior Vice President and General Manager, Compute at HPE.
"For these workloads, HPE ProLiant Gen13 brings together optimized performance, built-in, automatic security, intelligent operations, and proven reliability. It gives customers a trusted foundation to modernize and innovate today and adapt as their requirements evolve."
AMD described the launch as part of a longer-running collaboration between the two companies in enterprise compute.
"AI is moving into the applications and operations that run the enterprise, fundamentally changing what organizations demand from their infrastructure," said Dan McNamara, Senior Vice President and General Manager, Compute and Enterprise AI at AMD.
"Across generations of AMD EPYC CPUs and HPE ProLiant servers, our teams have worked side by side to turn advances in processor and system design into practical value for customers. With 6th Gen AMD EPYC processors and HPE ProLiant Gen13, we are delivering the performance, efficiency and flexibility enterprises need to scale AI while modernizing the critical workloads their businesses depend on."
Security push
Alongside the hardware, HPE is introducing a new version of its Integrated Lights-Out management technology, iLO 8. The release adds automatic self-encrypting drive protection from first boot, multi-person authorisation for high-impact actions and post-quantum cryptography readiness that extends beyond firmware.
HPE said the systems also include component verification tied to its Silicon Root of Trust approach. The feature is intended to check the authenticity and integrity of key server components as they are added, replaced or changed over the life of the machine.
On operations, Compute Ops Management now includes low-touch onboarding and near-real-time power capacity insight. HPE cited a commissioned Forrester study saying organisations using the management software spent 50% less time managing servers.
Release schedule
The four systems will not arrive at the same time. The DL525 is due later in 2026, the DL585a is scheduled for early 2027, and the XD245 and XD285 are set for release in 2027.
HPE said the Gen13 line will later expand to include more systems for mainstream enterprise, edge, software-defined infrastructure, database and AI workloads. According to the company, each ProLiant generation undergoes more than 1,000 tests and 10,000 hours of validation.