Rami Rahim’s message for network pros: Legacy networks can’t withstand rigors of AI

Last week’s HPE Discover event in Las Vegas marked the first time the annual user conference included the now-fully-absorbed Juniper Networks. Rami Rahim, former CEO of Juniper Networks and current EVP, president and general manager of HPE Networking, took to the main stage for a dedicated networking session after HPE CEO Antonio Neri’s morning keynote.

Rahim emphasized the network’s importance to AI and how AI is changing networking. He framed the AI conversation around one thing: network foundations. “AI is reshaping every part of the enterprise,” he said, “but none of that happens without the right foundation underneath it. And that foundation starts with the network.”

For network engineers, his keynote and customer conversations offered a clear blueprint for what comes next. Here are five key themes from the keynote and what they mean for your day‑to‑day work life and future career.

1. The network is now the AI foundation

Rahim opened with the story of San Francisco’s Millennium Tower, a visually stunning skyscraper that began to tilt because its foundation wasn’t built for long‑term environmental realities. He used it as an analogy for AI in the enterprise: Massive data movement, constant inference, real‑time responsiveness, and explosive scale will break any network that wasn’t designed for this new era.

“The network is no longer infrastructure sitting quietly in the background,” he said. “It’s become the strategic platform for how organizations operate, innovate, and scale.” You can buy “millions, if not billions, spent on GPUs,” he warned, but “if the network introduces latency and bottlenecks and instability, you’re limiting performance and slowing down outcomes.”

What this means for network engineers:

Architect for AI-era baselines, not legacy traffic assumptions. Design for sustained east-west flows, low jitter, and deterministic paths that support AI training and inference, not just classic north-south traffic.
Translate network health into AI business outcomes. When you speak to leadership, link latency, loss, and path diversity to model training time, inference SLA, or customer experience, not just “link utilization.”
Get in early on AI projects. Push to join the initial design discussions so connectivity, security, and data movement patterns are engineered in rather than patched later.

2. AI for networks: Self‑driving operations are becoming table stakes

The core thesis of Rahim’s presentation is that the legacy network can’t withstand the rigors of AI. “The old model of networking, static, manual, and reactive, simply cannot keep up with the speed and complexity AI introduces.” His alternative is an AI-native, self-driving operations model spanning Aruba Central and Mist, powered by Marvis, Marvis Minis, and an “agentic AI framework.”

On stage, he was joined by Sunalini Sankhavaram, VP of product management at HPE, who elaborated on this shift. “Experience-first AI in action” is built on “real-life experience data, every user, every minute, validated against real customer support cases, and enriched with digital twins.” In one demo, Marvis detected that “over 6% of user minutes were bad,” isolated the issue to a few overutilized APs, and “autonomously fixed the problem by enabling dual-band 5 GHz,” cutting peak utilization from 90% to 54%. Rahim summarized the outcome: “The network identified the issue, understood the root cause, determined the right action, and resolved the problem automatically before any user even had a chance to complain.”

For engineers, that’s a fundamental shift: from being the resolver to configuring, supervising, and governing these AI systems.

How to respond:

Lean into AIOps rather than treating it as a dashboard. Feed full-fidelity telemetry (“every user every minute”) into platforms like Mist and Aruba Central and actively test their recommendations.
Redefine your role as a guardrail designer. Your value increasingly lies in defining SLAs, allowed action scopes, and approval workflows for autonomous changes, rather than hand-tuning every parameter.
Learn the new language of operations: SLAs and “bad user minutes.” If your tooling speaks in experience metrics, you need to as well; they will become the common currency between IT and the business.

3. One AI‑native fabric across campus, branch, and routing

A big structural message in the keynote is unification: Juniper plus Aruba, Mist plus Central, wired plus wireless plus routing, all tied together by a common AI engine.

“We are innovating across both platforms to deliver a consistent self-driving experience,” Sankhavaram said. “With microservices, we can develop self-driving innovations once and deploy them on both the HPE Aruba Central and HPE Mist platforms,” much like a single app running on both iOS and Android. Marvis is being brought “right into the global north view” in Aruba Central, including the “Marvis Trust List”—actions you can set to run fully autonomously, such as automatically recovering a dead camera port to restore video without human intervention.

On the hardware front, HPE has already shipped a dual-platform AP and is bringing the “world-class HPE networking CX portfolio” to Mist for day zero, day one, and day two operations.” Rahim articulated the design principle clearly: “Our mission is simple. Bring the best innovations to both platforms, so that every customer in every industry gets the same powerful self-driving network, no matter which platform they choose.”

Advice for engineers:

Design for platform optionality. Assume that over the life of your hardware, management planes may change. Favor equipment that can switch between Mist and Central (or similar ecosystems) without rip-and-replace.
Embed digital twins into your workflows. Experience twins and synthetic testing should be part of pre-deployment validation and change management, not just a vendor demo.
Build API-first automation muscle. Sankhavaram emphasized “an API-first approach,” making data and actions accessible programmatically. That’s your cue to deepen your skills in Python, CI/CD, and infrastructure-as-code.

4. Networking and security are converging, with AI‑aware controls

Rahim repeatedly stressed that networking and security can no longer operate separately. “Attackers are already using the network as their weapon of choice,” he said, “and with AI making threats faster, smarter, and more sophisticated, defenders need to use the network as part of their defense.” This aligns with my research, which has found that 83% of network engineers now have security as part of their remit.

Customer voices underscored the point. “Security for us is job number one,” said Royal Bank of Canada’s Marlon Drummond. “We don’t have any other job other than protecting our client data. It’s our competitive edge, so we protect it with everything in our fiber.” RBC troubleshoots “at the network layer – the only place that you can get some immutable evidence,” using SD-WAN and DPI to create a “persona or personality for a user” and treat deviations as anomalies.

On the product side, HPE announced a unified SASE orchestrator that combines its Edge Connect SD‑WAN with its SSE stack into a unified SASE orchestrator with a single console. During the session, HPE showed an “AI‑aware firewall” that distinguishes sanctioned, unsanctioned, and tolerated AI apps, enforcing fine‑grained guardrails on uploads, prompts, and keywords. As Rahim put it, this lets customers “see, govern, and protect how AI is being used across their entire organizations without slowing down their businesses.”

For network engineers:

Expect to own more of the zero‑trust and AI‑governance story. Policies like “block ChatGPT, tolerate Gemini with upload and keyword controls” will be implemented in your SASE and firewall fabric.
Instrument the network as a primary security sensor. Follow RBC’s lead: build detection pipelines around network telemetry, lateral movement patterns, and per‑user “personas.”
Treat self‑driving changes as security‑sensitive. Autonomous routing shifts, RF changes, and port resets must respect segmentation and zero‑trust boundaries. Think like both a network engineer and a security architect.

5. Experience‑first networking at real‑world scale

The most compelling parts of the keynote were the customer segments, which showcased just how unforgiving modern environments have become.

Ohio State CIO Rob Lowden described a campus that is “a small city” with 66,000 students, 8,500 faculty, 22,000 HPE APs, and a football stadium that can see “200,000 plus people” in and around the venue on game day. With that density, “we need AI ops to crunch all of that data in real time,” he said. They’ve gone from problems that “can take hours” to being “resolved in minutes versus hours.”

At Sentara Health, director Tom Johnson did a great job highlighting the problem in his industry. “We move massive amounts of data across the organization, and it directly impacts patient care. That’s why our network must be resilient, secure, and always available. Because when data is delayed, care is delayed.” Ambient AI that listens to patient conversations and generates clinical notes is already in production, but “for those kinds of capabilities to work, our network has to deliver that data in real time, reliably, securely.”

Disney’s Ben Croy, who runs global networking, described studio networks where a single animated feature “could easily generate a petabyte of content,” with “over 200 concurrent productions globally.” In that world, the requirement is simple: “speed and simplicity… that’s it.” The network must be “foundational” yet “ideally invisible,” so filmmakers focus on story rather than connectivity.

What network engineers should take from these stories:

Measure what users feel, not just what devices report. Adopt SLAs around app quality (Zoom, EMR, VFX pipelines) as primary KPIs, and design your telemetry and AI‑ops to optimize those.
Use AI‑driven twins and synthetic tests to avoid “disasters that never happened” becoming real ones. Pre‑test big launches (stadium events, AI rollouts, global premieres) with synthetic users and paths, then let Marvis‑style systems validate and remediate.
Become a vertical expert. Whether you support healthcare, finance, education, or media, understanding domain‑specific workflows lets you prioritize and tune the network for what matters to clinicians, traders, students, or artists.

Final thoughts

Rahim closed with a challenge: “The old way of operating networks has reached its limits. The scale is too large, the complexity is too high, the pace of change is too fast, and AI is accelerating all of it.” In that world, “self‑driving networks are not a futuristic idea but a practical necessity.” For network engineers, the opportunity is to be the ones who build, govern, and extend that self‑driving foundation—before someone else does it for you.

I know many engineers still fear AI, but as the expression goes: AI won’t take your job, but another person who uses AI will.

Rami Rahim’s message for network pros: Legacy networks can’t withstand rigors of AI - Major Digest