HPE launches ProLiant server with NVIDIA Vera CPUs

HPE has introduced the ProLiant Compute DL394 Gen12 server with NVIDIA Vera CPUs, aimed at agentic AI workloads and large-scale data processing.

The new 2U server extends HPE's ProLiant line into a segment focused on reinforcement learning, sequential logic and other compute-intensive AI tasks that require strong single-core performance and high memory bandwidth. It is designed for organisations moving beyond generative AI models to agentic systems that need real-time reasoning and more predictable processing behaviour.

The launch is also part of a broader collaboration involving NVIDIA and Redpanda that the New York Stock Exchange is exploring. The exchange is examining the combination for market infrastructure that handles heavy message volumes and demands low latency and reliability.

At the centre of the server is NVIDIA's Vera CPU, intended to address AI and data processing workloads at AI-factory scale. The design aims to deliver high CPU performance, memory bandwidth and low latency for applications that need to process large volumes of data quickly and consistently.

Memory focus

A key part of the system's design is its approach to memory. HPE said the server uses NVIDIA Vera CPUs with a monolithic design rather than traditional high-core-count chiplet architectures, which it argued can create non-uniform memory access issues in multi-processor systems and lead to variable latency.

The server uses LPDDR5X memory and, according to HPE, reaches aggregate bandwidth of 1.2 TB/s, or up to 14 GB/s per core. HPE said this allows the system to ingest and process data at high speed, helps the CPU act as an orchestrator for demanding AI workloads and makes resource allocation more efficient.

HPE is also emphasising security and system management as part of the product's positioning. The DL394 Gen12 includes Silicon Root of Trust, the company's firmware security technology, alongside iLO 7 and a secure enclave designed to protect the server throughout its lifecycle.

According to HPE, these servers are the first to meet NIST quantum-computing-resistant security requirements. That, HPE said, makes the platform suitable for sensitive workloads and regulated environments where security controls are a key buying factor.

Management layer

The server also includes HPE Compute Ops Management, a platform that gives customers a single dashboard for overseeing distributed server environments. HPE said the software provides AI-driven operations intended to reduce management time and lower the risk of downtime.

Antonio Neri, President and Chief Executive Officer of HPE, said changing AI workloads are reshaping infrastructure requirements. "The shift from generative models to agentic systems is redefining the role of compute across the enterprise," he said.

"These workloads require high-performance servers with exceptional CPU performance to enable real-time reasoning across agentic AI and financial services applications. With our new HPE ProLiant Compute server, we are delivering a new class of infrastructure to help customers accelerate insights and operate with confidence in the most demanding environments," Neri added.

NVIDIA used the launch to underline its view that AI systems are moving towards architectures built for orchestration and coordination as well as model execution. "Agentic AI has arrived, and it needs a new CPU," said Jensen Huang, Founder and Chief Executive Officer of NVIDIA.

"Vera was built to orchestrate AI factories-delivering 2x the efficiency and faster task completion than x86. With HPE ProLiant Compute DL394 Gen12, enterprises can put Vera to work, and NYSE shows what purpose-built AI infrastructure can do in the world's most demanding environments," Huang added.

The New York Stock Exchange gave a more concrete indication of the deployment under consideration. Lynn Martin, President of NYSE Group, said the exchange is focused on system efficiency and resilience as trading infrastructure grows more data-intensive.

"At the NYSE, our focus is to optimize the latency, throughput, and reliability of the systems underpinning our unrivaled infrastructure," Martin said. "NYSE processes more than 1.1 trillion messages per day, and in collaboration with Redpanda and HPE, using NVIDIA Vera CPUs, we will be scaling our capacity while further optimizing latency to power a high-performance, resilient, and AI-ready market infrastructure."

HPE said the ProLiant Compute DL394 Gen12 will join its NVIDIA AI Computing by HPE portfolio when it becomes available later this year. The company is also offering the system through a financing programme that defers payments during the opening period of a deployment.