Oracle taps HPE Juniper for multi-year AI data center buildout

Oracle has signed a multi-year agreement with Hewlett Packard Enterprise to deploy Juniper networking equipment across its AI data centers worldwide, the companies announced Wednesday.

The deal covers HPE's Juniper PTX and MX routing platforms along with QFX and EX switching equipment, plus multi-year support services and financing. As part of the arrangement, HPE issued Oracle warrants to purchase shares of HPE common stock, though financial terms were not disclosed.

HPE Juniper gear already powers portions of Oracle Cloud Infrastructure's data center and edge networks. The expanded partnership specifically targets Oracle's AI supercluster expansion, where massive GPU arrays create intense demands on bandwidth, latency management, and fault tolerance.

Technical requirements for AI at scale

The latest QFX switches will deliver high-density connectivity and advanced congestion management for large AI backend networks using RoCEv2 (RDMA over Converged Ethernet version 2). The two companies are also developing intelligent telemetry tools designed to give Oracle earlier detection of queue buildup, packet loss, and component degradation—critical for maintaining GPU utilization and network uptime.

"Building AI infrastructure at Oracle's scale requires an integrated ecosystem where every layer of the stack works together," said Mahesh Thiagarajan, executive vice president of Oracle Cloud Infrastructure. "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."

Why it matters

Oracle's aggressive AI infrastructure spending—$55.7 billion in capital expenditures in fiscal 2026, up from $21.2 billion the prior year—reflects the massive physical buildout required to compete in cloud AI services. The company has guided toward roughly $70 billion in net capital spending for fiscal 2027. This HPE partnership provides Oracle with proven networking architecture capable of handling the extreme performance requirements of large-scale AI training and inference workloads, where even minor network bottlenecks can idle expensive GPU resources.

Networking revenue surge

The partnership announcement coincided with HPE's third-quarter earnings, in which the company raised its fiscal 2026 revenue growth forecast to 34% to 37% after posting record quarterly sales. HPE's networking revenue jumped 75% to $2.9 billion in the quarter, with data center networking more than doubling and routing revenue surging 270%.

Oracle's cloud infrastructure segment expanded 77% in fiscal 2026, driven by enterprise demand for AI capabilities and cloud migration. The networking infrastructure investment represents a foundational layer for Oracle's ability to deliver competitive AI services at the scale required by large enterprise customers.

Details of the partnership were first reported by Quartz.