HPE and Oracle Expand Alliance to Scale Global AI Data Center Infrastructure
Hewlett Packard Enterprise has announced an expanded, multi-year collaboration with Oracle to scale Oracle’s global artificial intelligence infrastructure by deploying HPE Juniper Networking platforms across Oracle Cloud Infrastructure’s (OCI) AI data centres, regional backbones, and edge network points of presence (PoPs).
Building on more than a decade of joint engineering between Oracle and Juniper Networks, the agreement encompasses the deployment of routing and switching platforms alongside multi-year networking support services and customised financing capabilities.
As Oracle accelerates the buildout of its massive AI superclusters, large GPU deployments are placing unprecedented pressure on underlying network fabrics. Sustaining optimal AI training and inference throughput requires predictable bandwidth, ultra-low latency, dynamic load balancing, proactive congestion management, and rapid fault recovery.
The integration of HPE Juniper Networking establishes a unified routing and switching architecture across OCI's distributed cloud footprint.
Multi-Platform Deployment: PTX, MX, QFX, and EX Series
The expanded agreement outlines the potential multi-year rollout of specialised HPE Juniper Networking hardware families across OCI's global data centre zones:
- PTX and MX Routing Series: Deployed across OCI's WAN and edge network layers, delivering high-density core routing, multi-terabit edge interconnectivity, and elastic bandwidth management to link regional clusters with external enterprise customer networks worldwide.
- QFX Switching Series: Powering multiple architectural layers of OCI's regional cloud data centre fabric. Advanced QFX platforms are deployed within high-throughput RoCEv2 (RDMA over Converged Ethernet version 2) backend AI clusters, providing line-rate switching, advanced Buffer Management, and packet-loss mitigation critical for distributed LLM training.
- EX Switching Series: Supporting resilient management, out-of-band networks, and enterprise-grade connectivity across data centre infrastructure bays.
Intelligent Fabric Telemetry for Maximised GPU Utilisation
In distributed AI training environments, a single degraded switch port, micro-burst congestion event, or dropped packet can stall thousands of GPUs executing synchronised parallel operations.
To safeguard cluster efficiency, Oracle and HPE are co-engineering intelligent network telemetry solutions:
- Early Queue Buildup Detection: Continuous monitoring of port buffer queues to trigger dynamic load balancing before congestion causes packet loss.
- Fabric Imbalance Analytics: Real-time visibility into traffic distribution across multi-stage Clos topologies to prevent hot-spot formation.
- Predictive Component Health: Automated identification of optical transceiver and physical link degradation to enable non-disruptive, proactive rerouting.
Rami Rahim, Executive VP, President and GM, Networking at HPE, highlighted the scale of the expanded partnership:
“Oracle is building AI infrastructure at extraordinary scale, and HPE is uniquely positioned to support OCI with one of the industry’s most comprehensive networking portfolios for AI data centers. Together, we are advancing a high-performance networking architecture that helps Oracle address the traffic management, power, space, and operational requirements of increasingly large AI environments.”
Mahesh Thiagarajan, Executive VP, Oracle Cloud Infrastructure, underscored the necessity of an integrated full-stack approach:
“Building AI infrastructure at Oracle’s scale requires an integrated ecosystem where every layer of the stack works together. HPE Juniper Networking has been an important part of our cloud architecture and now plays a key role in our AI clusters, regional data centers, and edge networks with the scale, performance, and reliability our customers expect.”
Praveen Jain, Senior VP and GM, Data Centre and AI Networking at HPE, emphasised the operational value of microsecond-level precision:
“Every microsecond matters to AI models and agents. Through high-performance networking and our work with Oracle on deeper fabric visibility, we are helping OCI detect bottlenecks faster, maintain network availability, and improve the efficiency of AI infrastructure.”
By combining HPE Juniper Networking's high-density routing and switching platforms with deep telemetry co-engineering, HPE and Oracle are building a resilient, high-bandwidth foundation capable of powering next-generation frontier AI models, agentic workflows, and global cloud services at hyperscale.
Read More:
2 Min read