SUSE on AI-Native, Sovereign Telecom Infrastructure

Samad Masood:

Hi, I am at Mobile World Congress 2026 and I'm joined by Richard and Rhys from SUSE. And we're going to talk about the frontier telco architecture, but also SUSE's approach to autonomous infrastructure.

Richard Card:

We're seeing a lot more conversation around digital sovereignty this year and how AI plays into that as well. Last year much of the conversation was around the infrastructure readiness for AI, whereas this year we're starting to see actual AI use cases running in production. So in our booth, for example, we're doing an AI demo with Infosys around the efficiencies of RAN so that's quite interesting.

Rhys Oxenham:

The Frontier Telco model is actually a really great way of describing some of the challenges and the opportunities that we're starting to see in the world of AI. I think that AI is of course changing fundamentally the way that we see how we can leverage optimizations, in not just technology, but also in how we can go out to market. And if you look at the way it's sort of broken down into, you've got the transactional, the operational, strategic layers of Frontier Telco. We can definitely start to see how organizations can really adapt the way the underlying infrastructure can be enhanced. And that frees up time and space for innovation to really use some of the human space to really differentiate as a business organization.

Richard Card:

The hot topic this year still continues to be AI, but it's also being driven by the need for digital sovereignty. And it isn't just restricted to Europe, but I think from some of the conversations I've had with CSPs, they see digital sovereignty as their opportunity to monetize the network, which is possibly something they've struggled to do previously with 5G. So for us we're positioning ourselves as sovereign by design. We're a European-based company, very much the ethos behind our business is sovereign-based. So that with our SUSE Telco Cloud proposition allows us to provide an infrastructure environment that telcos can continue to consume, A, to stay within the confines of the regulatory commitments of being sovereign, but also help them go and innovate and identify some of those new monetization routes to market.

Rhys Oxenham:

At SUSE, we see AI through two fundamental lenses. The first one is called infrastructure for AI and the second is AI for infrastructure. Let me talk about AI for infrastructure as kind of the main bucket here. This is where we are starting to see that really a paradigm shift has happened when you look at how you actually manage your infrastructure. So when you look at how do we identify faults, how we do remediation of faults, AI can actually start to go from actually autonomously making or remediating faults completely automatically. So when you look at the Frontier Telco model, when you actually look at ways and means by which the underlying infrastructure can start to be transformed through the works of agentics, you actually shift some of the actual responsibility for making those changes to the infrastructure itself. The humans that would have been involved in those leaps historically, they can be moved on to bespoke innovation, driving customer value in other areas where AI is now actually able to really boost that velocity and implementation speed.

The other one is of course, infrastructure for AI. And this is where SUSE is responsible for delivering on top of the, the Kubernetes stack that Richard was just talking about where we try and provide the best possible place to run AI workloads. What we try and do is we try and specialize in the areas where we have domain expertise. Of course, for us, we have a rich history in edge computing, but of course in telco as well, it's really trying to seek out opportunities where we can accelerate customer workloads and customer opportunities and business outcomes through the use of AI.

Specifically if we look at telco, there's a few different opportunities that we see for AI to accelerate some business outcomes. One of the big ones is of course energy optimization. If you look at the ways and means in which AI is able to analyze traffic patterns, the rate of which subscribers are coming online, maybe there's more density happening during a particular event that is happening. The infrastructure can effectively self-heal and be a little bit more optimized around, OK, during these times these cell towers can be maybe turned off. There's also in telco, not strictly related to the network but also core or customer-related activities, be that assistance with, billing, faults, these sorts of things. So there's plenty of opportunities specifically in telco, but a whole host of other use cases as well.

I think fundamentally in the industry we've really shifted from this world of 'OK, AI is here now, what are we actually doing with it?' We're kind of past that. I think historically it's been dominated a lot by the kind of shock and awe of generative AI, and it's now moving into, OK, how do we actually start deploying the infrastructure? How do we secure it? How do we scale it? I think that we're really starting to see where you can actually derive real business value. We're starting to see real tangible outcomes through the use and the application of AI. And I think that the types of practitioners, you know, especially if we look at the world of edge computing specifically, there's so many different practitioners that are involved in different areas of the stack. And I can see that AI will really start to become so much more pervasive throughout the infrastructure and especially when you look at agentics and it being able to take care of some of these responsibilities that perhaps a human would have done in the past.

There's a lot of view that AI is going to be incredibly disruptive in many ways it will. But also I see a huge amount of opportunity in that. Yes, AI will help accelerate business outcomes, it can help us with optimizations across a wide variety of use cases. But, I think the role for us as human operators and the bosses of those models. I think that the role of them is slightly shifting. I think it allows them to become a little bit more specialized in terms of their roles and what they have to focus on, allows them to really focus on value differentiation and a bit more innovation as opposed to doing maybe some more of the mundane tasks.

Richard Card:

I think the key areas for edge for telcos in terms of monetization is the enterprise space. Every telco is selling into all of the key verticals, that's multiple deployments. If you look at retail, thousands of stores for any particular chain. If we look at industrial IoT, for example, a significant number of endpoints. So for telcos the enterprise edge environment is the way to monetization.

Rhys Oxenham:

When we talk digital sovereignty, I'd like to use the word resilience as a digital resilience. And for many of edge organizations that are deploying infrastructure at the edge, they really need to think about how do I deliver service to my customers with autonomy and resilience. I cannot be dependent on third-party infrastructure or maybe cloud-based services. Look at a retail store, they need to be able to transact without that central infrastructure being there. And so many of the implementations that we're building out as part of our solution, they're not just sovereign by design in terms of the technology footprint and where they're built, but fundamentally designed to behave in some of the world's most hostile and threatening landscapes. And so whether that's through vandalism or through new security threats, or just simply due to physical proximity. Our software is built to be incredibly resilient. So, yeah, It's kind of in our DNA, in both the technology but also our heritage.

Richard Card:

Autonomy plays a key role in that. I mean, as Rhys pointed out, harsh environments, but if you have to roll a truck to do a software upgrade and it's 7 hours there, 7 hours back, or you have to send a helicopter, again economically it's inefficient. So if we can do these things with minimum human intervention, the profits start to go up.

Rhys Oxenham:

And that's probably a really strong tie in into the, the frontier telecom model where they're kind of advocating to Infosys, you're advocating for, you know, kind of a framing of this as well, maybe agentic AI and automated, you know, fault tracking remediation. This actually helps us from an edge model and the infrastructure becoming a lot more resilient because when you go to roll that truck, it's incredibly expensive to get someone on site. How can AI actually help with the ultimate remediation of it? It's a huge benefit.