This week, HPE customers and partners are gathering in Las Vegas for HPE Discover 2026, the flagship event focused on networking, cloud, and AI. For partners, the week began on Monday with the Partner Growth Summit. CEO Antonio Neri’s keynote had a telling title: “Building AI Starts with Your Network.”
In fact, over the past few months, it seems as though every conversation with a client revolves around the same thing. New models, new use cases, an endless stream of demos. But the more I work on it, the clearer it becomes that AI itself is rarely the real problem. The challenge almost always lies elsewhere.
Beyond the Experiment
Many organizations have now moved past the initial hype. Most have conducted a proof of concept (POC), built a chatbot, or tested their first GenAI use case. But as soon as you try to take that next step toward production, things start to get tricky.
Suddenly, questions start popping up. How do you get your data in order? How do you keep this affordable? What about security and governance? And how do you scale this without having to rebuild everything from scratch? This automatically leads you to something that sounds a little less sexy than AI itself: architecture.
Back to basics, but with a twist
What strikes me—even in HPE’s vision—is that the conversation keeps coming back to the fundamentals. Because there’s simply no other way.
- The network is once again taking on a central role. We’ve invested in applications and the cloud for years, but with AI, that network is once again becoming mission-critical. Data must move faster, more securely, and more consistently between the edge and the data center—in both directions. If you want to run AI at scale, your network has to be able to handle it. No bottlenecks, no surprises. You’ll notice the difference right away: AI that remains fast and affordable, even as usage increases.
- Hybrid is the reality—it’s no longer a choice. In theory, everything in the cloud still sounds great. In practice, I see very few customers who are actually there. Data remains scattered. Workloads, too. And AI just adds to that. So the question is no longer where you run your workloads, but how you keep the whole thing manageable. That’s where I see the real value of a platform like GreenLake: it brings complexity back under control. The result is an environment that you control, rather than one that’s chasing you.
- Data remains the limiting factor. Everyone wants AI. But not everyone has data that’s ready for AI. That’s probably the biggest difference between a cool demo and a working solution: access to the right data, in the right context, with the right governance. Once you get that in order, AI becomes reliable enough to really build on.
- Security suddenly becomes much more tangible. As long as AI merely makes suggestions, like a co-pilot does, it’s not too bad. But as soon as systems start taking actions on their own, the game changes. We’re moving toward AI agents that make their own decisions and execute processes. That’s powerful—and risky at the same time. Security then becomes something that’s built into the architecture from day one, not an afterthought. That way, you maintain control when systems start acting on their own, rather than having to intervene after the fact.
AI agents: fascinating and unsettling at the same time
What I personally find most fascinating is the shift toward agentic AI. These systems no longer merely provide support; they act on their own. They reason, make decisions, and take action. They’re a kind of digital workforce.
That sounds impressive, and it is. But at the same time, it highlights how many organizations aren’t ready for this yet. Because if an AI agent makes a wrong decision, who is responsible? And how do you intervene?
What does this mean for Xylos as a partner, and for you?
In my role, I’ve noticed that this fundamentally changes the conversation. It’s less about which AI use case we can build, and much more about whether the client is ready to actually implement AI.
That means thinking critically about architecture, making decisions about platforms, fostering open discussion about governance, and setting realistic expectations. And perhaps most importantly: having the courage to say that AI isn’t always the first step.
For your organization, it boils down to the same shift. The question isn’t which AI application you’ll build first, but whether your foundation is ready to support one. If you take an honest look at this, you’ll avoid costly detours and deploy AI where it truly pays off.
Conclusion: Less hype, more substance
AI is undoubtedly still a game-changer. Its real impact lies in what it forces us to do: organizations must get their fundamentals in order—networking, data, computing, and security. The basics, but at a level we didn’t need before.
Organizations that have that foundation in place move more quickly into production and keep their costs under control. They can also deploy agentic AI with confidence, while others get stuck at the stage of impressive demos. It’s less visible work than a new AI project, but it determines whether that project will ever deliver value. That’s where the real work lies for me, and where we, as a partner, can make a difference.
I’m curious to know how you experience this with clients or internally. Do you mainly run into use-case challenges, or are the underlying architecture challenges more common?
About the author
Frank Dierckx is a Partner Alliance Manager at Xylos and tracks developments in infrastructure, partner ecosystems, and emerging technologies. His expertise helps clients make technology choices that are technically sound and economically sound.
