HPE is expanding its self-driving network strategy. At HPE Discover in Las Vegas, the company unveiled new AI data center switches, additional AIOps features, and an integrated SASE platform. Innovations from the (former) Juniper portfolio are being incorporated into the HPE AI Factory. Juniper Mist is being integrated into Aruba CX switches, enabling them to become self-driving and reducing the workload for network administrators.
With these announcements, HPE is positioning the network as the foundation for what it calls the “agentic enterprise”: environments in which AI agents perform tasks independently. According to the company, this requires networks that monitor themselves, predict outages, and resolve them autonomously, without a network administrator having to intervene every time. Rami Rahim, head of the networking division, states that the success of agentic AI stands or falls on a modern network foundation, in which performance, reliability, and intelligence determine the effectiveness of the entire AI architecture.
HPE Networking brings together the best of Aruba and Juniper. The Mist AIOps platform will now support Aruba CX switches, expand the Marvis AI in Aruba Central with self-healing capabilities, and introduce new AIOps features that use “agentic reasoning” to identify the root cause of outages more quickly.
Also read: HPE Networking firmly takes the wheel of the self-driving network
Switches for inferencing and scale-up
The biggest step forward is in AI networking. HPE is incorporating its Juniper networking division into the so-called AI Factory validated design, a pre-integrated stack of compute, networking, storage, software, and services. The Juniper QFX switches are managed via HPE Networking Data Center Director. With this, HPE is targeting customers who want to move their AI infrastructure from proof of concept to production.
Two new switches round out the portfolio. The QFX5140 is designed for inference clusters and AI at the edge, a rapidly growing market. The QFX5252 is a scale-up module (switch tray) designed specifically for the AMD Helios rack, optimized for low latency and high bandwidth. The goal remains the same: to reduce the time GPUs spend waiting on the network, allowing them to devote more time to computing and lowering the total cost of ownership.
AIOps rolled out more widely
HPE is further extending its AIOps capabilities across its entire network portfolio and bringing the Aruba Central and Mist platforms closer together. Through Mist, the CX switches will gain features such as zero-touch provisioning, wired assurance, and Marvis-driven actions. In Aruba Central, Marvis will introduce self-driving capabilities.
In the data center, HPE is adding two new features. Predictive maintenance uses AI and machine learning to predict system failures before they occur. Additionally, there is a “reasoning agent” that can autonomously search through diverse data sources, ranging from millions of support cases to a contextual graph database, to determine the root cause of issues and propose concrete remediation actions.
Furthermore, Mist Networking Data Center Assurance is now also integrated into Compute Ops Management and within GreenLake. This is intended to reduce the proliferation of management tools.
A single console for network and security
HPE is also introducing a unified SASE platform, built on EdgeConnect. It brings SD-WAN and cloud security together in a single AI-native console. The company points out that attackers are using AI to find vulnerabilities faster and positions the platform as a way to accelerate the adoption of zero trust. A built-in SSE connector enables this without additional infrastructure, while a Private Edge variant keeps traffic within the company’s boundaries. This is intended as the foundation for “sovereign SASE.” Management is partially handled via natural language using a SASE Copilot.
All in all, HPE is making significant strides in networking. The acquisition of Juniper is beginning to show results within HPE Networking, but this appears to be just the beginning.