By Gary Mackenzie, GM of Telco Business at SUSE
This will be my tenth Mobile World Congress. That gives you a certain perspective. MWC has always been an industry constant – through technology shifts, market cycles, and even a global pandemic. It has never really chased fashion. It’s where long-term bets surface, where operators test assumptions, and where reality has a habit of catching up with rhetoric.
That’s why I’m genuinely looking forward to Barcelona this year. The conversations already feel purposeful. There’s energy, but it’s a grounded kind – less about what might be possible one day, and more about what needs to work, everywhere, now.
AI will be everywhere again, of course. But what’s striking this year isn’t its visibility, it’s the tone. The breathlessness has eased. AI is no longer a novelty to be admired, but an asset to be managed. And once something becomes assumed, the industry starts asking harder, more interesting questions.
Those questions are familiar ones: where does this run, how do we operate it, how do we secure it, how do we scale it, and who is accountable when it doesn’t behave as expected? These are infrastructure questions, and infrastructure has been living with them for decades.
In that sense, AI and infrastructure have reached the same point at the same time. Infrastructure went through its own reckoning after years of abstraction promising simplicity and speed. Eventually, gravity returned. Power, latency, lifecycle management, compliance, and cost all reasserted themselves. AI is now arriving at that same moment – not collapsing under hype, but settling into delivery reality.
That shift is healthy. The optimism hasn’t gone away; it’s simply become more disciplined.
For a long time, progress in our industry was measured by what could be demonstrated. Proofs of concept mattered. Pilots mattered. They still do, but they’re no longer differentiators on their own. What matters now is whether platforms can be run, upgraded, and evolved over time. Whether they can be secured without heroics, automated without fragility, and operated by real teams under real constraints. This is the unglamorous center of modern engineering. It doesn’t make for dramatic keynotes, but it’s where outcomes are decided.
One of the most encouraging shifts I’m seeing is a more honest discussion about scale. Nothing suddenly breaks when environments grow—behavior changes. What feels clean and manageable at five clusters becomes fragile at five hundred. Manual interventions turn into systemic risk. Complexity compounds.
AI workloads amplify this effect. Data locality matters more. Scheduling pressure increases. Failure domains widen. The infrastructure beneath AI stops being a background concern and starts shaping results directly. This isn’t about limiting ambition; it’s about recognising that repeatability, automation and operational discipline matter more than architectural elegance as systems grow.
Telecommunications has always forced these lessons early. Scale arrives fast. Distribution is unavoidable. Downtime is visible and expensive. That’s why telco environments tend to expose weaknesses long before other sectors do.
What’s particularly interesting now is how operations are converging across edge, core and IT environments. Not into a single architecture – that’s rarely realistic – but into shared operational models. Standardised blueprints. Declarative infrastructure. Converged lifecycle management. The focus is shifting from bespoke silos to operations-ready platforms that can be applied consistently, whether you’re running a handful of sites or thousands.
Alongside this, sovereignty has become a practical concern rather than a theoretical one. It’s no longer just about data residency as a checkbox. It’s about control: over software supply chains, over operational dependencies, over where decisions are made and enforced. AI accelerates this conversation rather than creating it. When workloads are critical, distributed, and data-intensive, sovereignty becomes an operational requirement, not a policy footnote.
In 2026, “boring” is the highest form of engineering maturity. It’s a compliment. Boring upgrades. Boring rollbacks. Boring observability. Systems that behave as expected, even under stress. Failure modes that are understood rather than surprising. This isn’t stagnation, it’s what allows innovation to happen safely.
At SUSE, this is how we think about the next phase of telco cloud and AI infrastructure: open by design, sovereign by choice, and built for delivery. Open standards to avoid lock-in. Declarative operations to eliminate drift. Industry blueprints to reduce integration cost. And a platform that scales operationally, not just architecturally.
That’s why I’m looking forward to MWC this year. Not for the keynotes – the real signal is never there – but for the conversations in meeting rooms, on stands, and while walking the halls. The repeated themes. The quiet consensus forming around what really matters next.
After ten MWCs, what still excites me is that moment when an industry stops asking what’s possible and starts agreeing on what needs to work.
See These Ideas in Action at MWC
Many of the themes above are reflected in the demonstrations and customer stories we’ll be sharing in Barcelona. You can explore them in more detail below and at event at Hall 2, Stand 2C31.
(Visited 1 times, 1 visits today)
