HPE Discover Barcelona 2025 unveils AI networking advances
HPE Discover Barcelona 2025 highlights AI networking advancements, including the integration of HPE Aruba Networking Central and HPE Juniper Networking Mist into a dual platform. The event also introduced the HPE Juniper Networking QFX5250 switch and MX301 router, enhancing AI data center capabilities. HPE aims to unify AIOps platforms and deliver secure AI-native networks, focusing on AI for networks and networks for AI. The company collaborates with Broadcom to offer AMD Helios AI rack-scale architecture with integrated Ethernet networking.
HPE has been a busy bee with a push for "cross-pollination" between its AIOps platforms.
Months after closing its acquisition of Juniper, HPE announced it has simplified IT operations with HPE Aruba Networking Central and HPE Juniper Networking Mist by creating a dual platform. The ultimate goal is to unify the two AIOps platforms through what an HPE executive refers to as "cross-pollination" of capabilities, enabled by microservices, from one platform onto another, and work toward a self-driving network.
"Dual platform design lets our customers choose their preferred control point, whether it is Mist or Aruba Central, and switch seamlessly between platforms as their needs evolve with no new hardware involved in that decision," said Rami Rahim, president and general manager of the HPE networking business, in a press briefing.
However, a true self-driving network that is intelligent enough to automatically adjust to various conditions to reduce performance degradations and outages is still far off but it is getting incrementally closer with integrated AI technologies, specifically agentic AI, according to Jim Frey, a principal analyst at Omdia.
"Both the Juniper and Aruba teams were making progress in this direction independently, and now they can put the best of their AI capabilities together to accelerate progress," said Frey.
This incremental sharing of existing AI capabilities across Aruba Central and Juniper Mist includes:
- Mist's Large Experience Model will now be available in Aruba Central.
- Aruba Networking's agentic mesh technology will be available for Mist, which will enhance anomaly detection and root-cause analysis.
- Mist will adopt the organizational insight and global Network Operations Center views from Aruba Central for a unified UX.
- New WiFi-7 access point models will work across Aruba Central and Juniper Mist, aiding buyer protection.
"It is encouraging to see these integrations, but the company has a long journey ahead of it. Ultimately, the company may have to commit to one platform in favor of the other in order to maintain its pace of innovation," said Shamus McGillicuddy, vice president of research at Enterprise Management Associates.
It's time for a switch
Rahim said HPE's objective is to deliver secure AI-native networks, and it is focusing on two key market opportunities: AI for networks and networks for AI.
A new addition is the HPE Juniper Networking QFX5250 switch, 100% liquid cooled, which connects GPUs within data centers with a high performance Ultra Ethernet Transport-ready switch. It is the first OEM switch built on Broadcom Tomahawk 6 silicon and touts a 102.4 Tbps bandwidth. It will be available in early 2026.
"The QFX switch puts HPE in a temporary leadership position for AI networking. I have no doubt that Cisco and Arista will announce comparable switches within the next few months, both based on the same Broadcom silicon," said McGillicuddy.
HPE also introduced the HPE Juniper Networking MX301, a multiservice edge router. The compact 1RU router delivers 1.6 Tbps performance and 400G connectivity across inference, multiservice, metro, mobile backhaul and enterprise routing environments. Rahim described it as "absolutely ideal as an on-ramp for the distributed inference cluster opportunity."
"For those building out new data centers or upgrading existing ones, [these products] now represent the highest performance and capacity for routing and core switching that [HPE] offers," Frey said.
In a push to meet industry demand for AI compute capacity, HPE is working with Broadcom to offer the first AMD Helios AI rack-scale architecture with integrated scale-up Ethernet networking.
"We're bringing Ethernet to a new layer of the AI data center network for the first time," Rahim said.
McGillicuddy commented that HPE’s development of an Ethernet scale-up switch for Helios will expand its overall position as a leading AI networking provider.
"The combination of this component with the new QFX and MX products means that customers have a broad set of networking products available to them as they build out AI data centers, all based on the same network operating system with a common tool for network management," McGillicuddy said.
Kathleen Casey is the site editor for SearchCloudComputing. She plans and oversees the site, and covers various cloud subjects including infrastructure management, development and security
