How do you build profitable AI-native networks from 5G-Advanced to 6G?

Discussions have begun on building profitable AI-native networks from 5G-Advanced to 6G, and the impact of AI on how we build radio access networks (RAN). TelecomTV brought together experts from Nokia and NVIDIA to explore how AI-RAN is reshaping mobile networks and discover what telecom providers need if they are to capture the business opportunity unlocked by AI-RAN.

The objective of the panel, which was moderated by TelecomTV’s Guy Daniels and included Mikko Jarva, head of portfolio and technology for network APIs at Nokia, Aji Ed, head of AI-RAN and cloud RAN at Nokia, and Soma Velayutham, vice president of AI and telecoms at NVIDIA, was to identify what telecom providers need to capture the business opportunity unlocked by AI-RAN. It also aimed to discuss the growing importance of programmable networks in enabling the integration of network capabilities with enterprise applications, creating opportunities for cross-industry innovation.

Ed cited NVIDIA’s landmark $1bn investment in Nokia, announced last October, as a turning point for the industry. “It brings together two global leaders, Nokia in wireless networks, and NVIDIA in AI and accelerated compute. But what is exciting is we are now able to bring GPU-accelerated compute right to the edge of the RAN network, to the base station, to the cell site,” he said. “Instead of sending everything back into the datacentre, we can run real-time AI in the baseband and we can, therefore, bring smarter algorithms, better spectral efficiency and the ability to improve the performance through software-defined architecture. That is the key.”

He added that this perfectly lays the path to AI-native 6G, where the network becomes intelligent, adaptive and capable of hosting entirely new applications and services. This is not just an investment; it is a catalyst. It accelerates the evolution from a network that connects people to one that connects intelligence, connecting people to intelligence, Ed explained.

The base stations of tomorrow will also be software-defined and programmable, with AI as a value add. This will supercharge edge AI. Many applications will be developed but, for now, it’s all about creating value with AI.

When asked how a GPU-powered platform could enable new revenue sources and what it meant for possible applications and monetisation opportunities for telecoms providers, Soma Velayutham said this is the first computer in the world capable of running high-performance RAN workloads like Nokia virtualised RAN (vRAN) and at the same time deliver the tokens for ChatGPT. “So you have one heterogeneous computer that can actually do RAN very well and, at the same time, do tokens very well. And you know that in parallel, there is a huge insatiable demand for tokens and delivering tokens in the world. You can see the growth of GPU-as-a-service, token-as-a-service, and model-as-a-service. I mean, this is just growing exponentially, depending on whose report you read, a hundred billion-dollar market. Now, every operator who is actually deploying the NVIDIA compute platform and the Nokia RAN software can monetise that compute asset instead of just using it 30% of the time on average and occasionally using it during the peak.”

Nokia’s Mikko Jarva pointed out that capabilities, such as AI compute at the edge, an AI-native network, programmability and the software-defined nature of the network design will be core to 6G. “The question is how do we utilise them and what are they used for? Programmability of the network leverages these capabilities to optimise the delivery of specific workflow data. Soma [Velayutham] was referring to the tokens, making sure that the first and last token get through the network as fast as possible into the right AI workload. That’s about optimising the network for AI workloads and other workloads, and the software-defined nature of the network enables that much better. It supercharges the capabilities of the network to deliver tokens. Of course, related to that is also optimising how and where the network data flows in the network. So it’s about optimising routing.”

The new applications are all about connecting intelligence and, if possible, through software-defined means, shape the network for connecting intelligence. That will then help power value creation. And of course, with value creation, monetisation follows, added Velayutham.

AI applications, mission-critical applications, whether for autonomy, security or first responders – it’s about video analytics, then inference or demand inference, high precision, high-fidelity inference on demand in a closed and secure place.

AI for RAN brings two types of improvements: One is operational efficiency improvement, and the second is spectral efficiency improvement. When referring to operational efficiency improvement, it means building AI-enabled automation and tooling like self-organising networks (SON), service management and orchestration (SMO) and all of these different tools that help to improve the operational efficiency of the network. GPU-accelerated computing with AI-RAN helps in spectral efficiency improvement, and spectrum is the biggest asset for an operator, explained Velayutham.

Everything being done in AI RAN, whether it’s 5G, 5G-Advanced or 6G, is all software defined. “This compute platform that operators will invest in and deploy is future-proof, and it’s also very flexible. They can choose the kind of fronthaul they need, whether they want it to be a different split, or they can choose different radio unit (RU) vendors, or they can choose different configurations. As C-RAN and D-RAN, it’s completely futureproof, and because it’s built with AI native, all the new algorithms, if you look at 3GPP Release 20 and 21, are all about a large number of AI-based algorithms, and this is the best platform to actually start delivering AI-based algorithms in the future,” he added.

The first transformation is making the network as a whole into an AI-native construct and a platform, and applying fully software-defined principles. Then the next layer is how that network becomes programmable and how it’s being accessed by applications, how the applications signal to the network what they need and how the applications bring back or get data from the network to understand the network’s capabilities.

Jarva called what the network sees and senses ‘AI-native exposure’. “That means AI native and agentic interface towards the network. And this would be obviously building on top of APIs through MCPs [Model Context Protocols] and also exposing network-specific agents as gateway protocols towards the consumption of the network. And then the third layer is the agent, the consumption of the network. So, we will see more and more external agents accessing network capabilities, placing compute into the network, delegating tasks for the networks to do and shaping the network for the mission that those agents are using. For these three layers, AI-native network holistically as the platform AI-native exposure, how we communicate and exchange data with that network, and then the consumption of the network through AI agents.”

Ed added, “When we talk about AI on RAN, AI for RAN, AI and RAN, we have real demonstrations, real use cases available. And at the same time, we are also expanding with more and more customers. We’ll be announcing more and more customer partnerships, and we will have performance and power-efficiency demonstrations and competitiveness, which we will execute during the course of this year. And on top of it, there are multiple different milestones we have defined in terms of the different proof of concepts and trials, which will validate all these different aspects of performance, spectral efficiency improvement, power efficiency and all of these different aspects in 2026. We are truly excited, and this is the kind of journey that we want to undertake for the future to create AI-native 6G.”

Related content: Building profitable AI-native networks from 5G-Advanced to 6G