The first HPE ProLiant server enabled with NVIDIA GH200 NVL2 — a 2U platform with up to two NVIDIA GH200 Grace Hopper Superchips and large coherent memory, purpose-built for AI inferencing on large language models.
Do you need to run inference on large language models that demand enormous memory capacity and bandwidth?
The HPE ProLiant Compute DL384 Gen12 is a 2U rack server enabled with the NVIDIA GH200 NVL2 platform, supporting up to two NVIDIA GH200 Grace Hopper Superchips. By tightly coupling NVIDIA Grace CPUs with integrated Hopper GPUs over a coherent, high-bandwidth memory fabric, the DL384 Gen12 gives large-language-model inferencing the memory capacity and bandwidth that a conventional GPU server cannot match — wrapped in the security, manageability, and lifecycle tooling enterprises trust from HPE ProLiant.
| Form factor | 2U rack; NVIDIA GH200 NVL2 platform |
|---|---|
| Superchips | Up to 2 NVIDIA GH200 Grace Hopper Superchips |
| CPU | NVIDIA Grace — 72 Arm Neoverse V2 cores per Grace CPU |
| GPU | Integrated NVIDIA Hopper GPU per superchip |
| Memory | LPDDR5X CPU memory + HBM3e GPU memory; up to ~1.2 TB coherent memory (dual-superchip) |
| Storage | Up to 8 EDSFF NVMe Gen5 (E3.S); 2 M.2 boot devices |
| Expansion | Up to 4 PCIe Gen5 x16 slots; up to 2 OCP 3.0 slots |
| Management & security | HPE iLO, Silicon Root of Trust |
Download datasheet: HPE ProLiant Compute DL384 Gen12 QuickSpecs (HPE.com)
The NVIDIA GH200 Grace Hopper architecture unifies CPU and GPU memory over a high-bandwidth coherent fabric, giving inference workloads far more addressable memory than a conventional discrete-GPU server. That makes the DL384 Gen12 well suited to large language models, LLM fine-tuning, retrieval-augmented generation, and simulation.
The HPE Silicon Root of Trust establishes a zero-trust security framework at the silicon level to ensure firmware integrity, continuously detecting compromised servers and preventing them from booting if malicious code is found — the same fundamental security approach engineered across the HPE ProLiant Compute Gen12 portfolio.
Manage the DL384 Gen12 with the same HPE tooling as the rest of your fleet, including HPE iLO and HPE Compute Ops Management for global visibility, automation, and simplified lifecycle management. Available as a standalone purchase or as a service via HPE GreenLake.
Tell us about your large-model AI workloads and a ServerComputeWorks specialist will help you spec the GH200 configuration, price it, and place the order.
Request a Quote*NVIDIA, Grace, Hopper, and GH200 are trademarks and/or registered trademarks of NVIDIA Corporation. Arm and Neoverse are trademarks of Arm Limited. All other third-party trademark(s) is/are property of their respective owner(s). Specifications and availability are subject to change.