Hewlett Packard Enterprise chief executive officer Antonio Neri took a moment during the bustle of the company’s Discover 2026 show this week in Las Vegas to look back to almost a year ago, when HPE finally closed its $14 billion acquisition of rival network vendor Juniper Networks after a bruising, year-plus long negotiation with the US Justice Department.

A decade before, HPE had bought Aruba Networks, giving it a networking play in branch and campus environments to complement what it offered in the datacenter. However, Neri said that in the years since, with enterprise cloud adoption continuing to expand and the rise of AI, “we learned very quickly that the next big opportunity, the next big frontier of innovation – the networking layer of the stack – is going to be the next opportunity.”

The company needed to strengthen its core networking portfolio to keep up with rivals like Cisco Systems and Arista Networks, and Juniper with its lineup of datacenter switches and routers seemed like a good match. The growth in AI inference workloads and the emergence of AI agents made such a deal even more important.

“In my mind, [the Juniper acquisition] was perfectly timed,” he said. “Now we have a complete edge to core to cloud portfolio, perfectly tuned for the AI transition inflection point that we see, and that's similar to the rest of the portfolio, whether it's in cloud, storage, and servers, or whether it is in AI at scale with compute.”

As we noted when the Juniper deal closed, the question became about how the product portfolios would be integrated. The answers began coming relatively quickly. At HPE Discover Barcelona 2025 in December, Rami Rahim, former chief executive officer of Juniper who now is executive vice president, president, and general manager of HPE Networking, said the idea is to use the combination of Mist AI technology inherited from Juniper with the Aruba Central platform to fuel HPE’s strategy of the self-driving network, a plan to embed AI throughout the network fabric – from the cloud to AI datacenters and out to the edge – to create a network that can proactively configure, optimize, and heal itself with little to now human intervention.

At Discover this week, HPE continued the effort with a series of integrations – not only within the networking portfolio, but also between networking and compute – and new products. Mist AI is in the middle of much of the integration. It includes following through on Rahim’s promise to integrate Mist into Aruba Central, and Aruba Center into Mist, a key part of what HPE calls its “cross-pollination” integration strategy to unite the Juniper and Aruba portfolios.

The move will allow the two platforms to share agentic capabilities and hardware and drive consistencies across AI-native operations. Fueling this integration is the Marvis AI engine – another technology from Juniper – that acts as an AI network assistant to help with automated troubleshooting of the network. Marvis Actions is an AI-powered component that pinpoints the cause of network issues and can autonomously resolve the issue.

HPE is making Marvis Actions available for Aruba Central and is integrating its CX switching portfolio with Mist. Later this year, Marvis Actions will also come to Aruba Central. In addition, HPE is adding predictive analytics to Mist so it can proactively predict system failures and prevent network outages, as well as the ability to use an advanced reasoning AI agent for remediation tasks, bringing root case analysis for the datacenter network.

“Think of this as Marvis AI Engine for datacenter operations,” Rahim said. “We're combining telemetry, application flows, operational context, historical knowledge to understand rapidly the root cause and recommend next steps. Problems that once took hours, if not days, to diagnose can now be resolved literally in minutes or even proactively before anybody understands that there is an issue.”

In addition, HPE is integrating Mist Networking Data Center Assurance – a suite of cloud-based AIOps software used to automate, optimize, and secure private cloud and datacenter networks – with both HPE Compute Ops Management and GreenLake. Juniper networking technology also is being integrated into HPE AI Factories.

On the security front, HPE introduced the SASE (Secure Access Service Edge) Orchestrator, bringing together SD-WAN and SSE (Security Service Edge).

New hardware includes the QFX5140 switch for AI inference cluster and AI jobs at the edge – a way of ensuring AI inferencing can be done where the data is:

There is also the new QFX5250 Switch tray for AMD’s “Helios” AI server rack to enhance the performance of AI infrastructure at scale:

Neri added in the meeting with journalists afterward that HPE has a “fantastic portfolio and that portfolio is equal or better than Cisco in many ways because we believe our architectures are modern, cloud-native and AI-driven.”

HPE isn’t the only vendor seeing the effect the rise of agentic AI and inference workloads will have on networks. At their Cisco Live 2026 event earlier this month, executives continued to build out its enterprise AI infrastructure stack to deal with emerging technologies like agentic AI and frontier AI models. Like their HPE counterparts, they stressed the foundational nature of networks in the stack and how they need to adapt to the growing agentic and inference workloads.

In an accompanying report about the impact of AI on wide area networks (WANs), Cisco noted that it goes beyond simply increasing traffic, though that is a concern. The report found that agents generate 450 percent more traffic per task than traditional tools do. That said, agents change the traffic itself, moving away from the human-paced video streams that traditional networks were expected to address.

“By 2035, one-quarter of network traffic is projected to be AI inference,” the report’s authors wrote. “These flows don’t behave like the web. They live longer, demand more upstream capacity, and operate at software speed, not human speed. The connectivity between agent logic and AI models effectively becomes the agent’s ‘spinal cord’ – a critical dependency whereby any network degradation directly impairs agent functionality.”

Neri said HPE understands the changes agentic AI and inferencing will bring to networks, and how important those networks are.

“There is always one core element of your infrastructure,” he said from the keynote stage. “With AI, that core element is the network. The performance of your entire architecture depends on it. Every byte, every token, every decision, all of it crosses the network, which is why today we are bringing the HPE Juniper network into our AI data solutions, enabling more efficient, high-performance AI environment. Whether you are a hyperscaler, service provider, or a neocloud, or a larger enterprise, you have more choice in how you connect and secure your largest AI investments.”