HPE Launches AI Grid for Distributed Edge Inference

Hewlett Packard Enterprise launched HPE AI Grid, a distributed infrastructure package designed to enable AI inference at the edge. Based on a reference architecture from NVIDIA Corporation, the system connects AI factories with far-edge inference sites to provide service providers with low-latency connectivity.

The technical stack integrates HPE ProLiant servers and Juniper routing assets with NVIDIA RTX PRO 6000 Blackwell GPUs, BlueField DPUs, and Spectrum-X Ethernet switches. Comcast has initiated field trials focusing on small language models and real-time edge AI inferencing, while TELUS and CityFibre have expressed interest in using the technology for network optimization and service delivery.

To encourage adoption, HPE Financial Services is providing lease savings for AI-ready networking and 0% financing on specific AIOps software. The collaboration aims to create an intelligent fabric for distributed inference by combining accelerated computing with telco-grade multicloud routing.