Hewlett Packard Enterprise (HPE) has launched its ProLiant Compute XD685 servers, specifically designed to support artificial intelligence (AI) clusters utilised in tasks such as training expansive language models. Aimed at addressing needs of large AI model builders, including government and enterprise clients, this server system is empowered by the latest fifth-generation EPYC processors and Instinct MI325X accelerators from Advanced Micro Devices (AMD). AMD revealed these processors at its recent Advancing AI event.
HPE's ProLiant Compute XD685 is currently open for orders and is expected to be generally available in Q1 2025. Bank of America analysts, ahead of the unveiling, underscored HPE's potential to profit from increasing demand for robust AI server racks, mainly due to its liquid cooling technology. As larger, higher-density data centers call for liquid cooling to manage their temperature, this could represent a significant opportunity for the company.
While HPE's shares dropped by 1.5% to $20.54 on Thursday, they have risen by over 20% for the year 2024.