HPE unveils new AI and Supercomputing solutions for enterprises and service providers

Hewlett Packard Enterprise (HPE) has introduced new high-performance computing (HPC) and artificial intelligence (AI) solutions, aimed at enhancing research and AI training for large enterprises and service providers. The portfolio includes advanced supercomputing systems optimized for tasks like large language model (LLM) training and natural language processing (NLP). 

The HPE Cray Supercomputing EX series is at the heart of this new portfolio, featuring cutting-edge direct liquid cooling technology to enhance efficiency. Among the key systems is the HPE Cray Supercomputing EX4252 Gen 2 Compute Blade, which offers up to 98,304 cores in a compact rack. This compute blade, powered by AMD’s 5th Gen EPYC processors, is set to become available in Spring 2025. Another significant system, the HPE Cray Supercomputing EX154n Accelerator Blade, is designed to speed up AI and HPC workloads by accommodating up to 224 NVIDIA Blackwell GPUs. It is expected to be available by the end of 2025. 

Additionally, the HPE Slingshot Interconnect 400 will offer network speeds of 400Gbps, designed to support large AI clusters with reduced network infrastructure needs. This interconnect will be available by Fall 2025. The HPE Cray Supercomputing Storage Systems E2000 will provide high-performance storage for HPC and AI workloads, with a launch planned for early 2025. 

To complement these systems, HPE is introducing the HPE ProLiant Compute XD servers, which are optimized for AI training and model deployment. The HPE ProLiant Compute XD680, featuring Intel Gaudi 3 AI accelerators, will be available in December 2024. The HPE ProLiant Compute XD685, which includes NVIDIA GPUs, will launch in early 2025, offering improved performance for demanding AI models. 

HPE also integrates its proprietary HPE Integrated Lights-Out (iLO) technology for enhanced security and streamlined management of AI clusters. With these new solutions, HPE aims to help organizations accelerate their AI deployments and drive faster scientific discoveries.  

“Service providers and nations investing in sovereign AI initiatives are increasingly turning to high-performance computing as the critical backbone enabling large-scale AI training that accelerates discovery and innovation,” said Trish Damkroger, senior vice president and general manager, HPC & AI Infrastructure Solutions at HPE.