Why 2026 Matters for 6G and Sustainable AI

FutureNet World 2026 had a strong focus on AI-driven operations, autonomous networks and the future of intelligent telecoms infrastructure. On Day 2, NGMN led a dedicated Track C session titled The Path to 6G & Sustainable AI, bringing together operator speakers from Orange, Deutsche Telekom, TIM and DOCOMO Euro-Labs, along with moderators from STL Partners and TelecomTV. The track covered 6G development, AI implications, network simplification, cloud-native evolution, agentic AI-based operating models and the future shape of the next generation network operator.

The opening fireside chat between Phil Laidler of STL Partners and Laurent Leboucher of Orange framed 2026 as a critical year for 6G. This does not mean that 6G is about to become a commercial reality. In fact, the discussion pointed towards 2030 as the more realistic timeframe for commercial deployments. The reason 2026 matters is that the foundations are being shaped now. 3GPP studies have started, specifications are expected to begin in 2027, and the architectural direction being set today will influence what operators, vendors and the wider ecosystem will be working with for the next decade.

The key difference this time is AI. Previous mobile generations largely followed a familiar rhythm: new radio, new spectrum, new hardware, new capabilities and a roughly ten-year industry cycle. AI does not work on that timescale. Its evolution is measured in months rather than years. That creates a major challenge for 6G, because the telecoms standards process needs to preserve global interoperability while also allowing faster innovation cycles around AI, automation, software and cloud-native platforms.

One of the most important messages from the discussion was that 6G cannot simply be another forced network refresh. Operators are still working through the realities of 5G monetisation, 5G standalone deployment, legacy complexity and operational cost. A new generation only makes sense if it brings clear value. The 6G discussion therefore needs to be demand-driven rather than technology-driven. It needs to focus on the capabilities that matter, avoid unnecessary over-specification, and support a smooth evolution from today’s 5G systems.

Kostas Chalkiotis from Deutsche Telekom then presented NGMN’s work on AI Surge and its Implications for 6G. NGMN’s publication examines three key areas: the impact of AI-driven traffic on networks, the network capabilities needed to support AI-based services, and the role of AI in future network architecture evolution.

A useful way to think about this is through the two familiar angles: AI for networks and networks for AI. AI for networks is about using AI to operate, optimise and secure networks more effectively. Network for AI is about how the network itself may need to evolve to support AI services, AI agents, distributed intelligence, changing traffic profiles and new requirements around latency, uplink, trust and security.

The important point is that there is still a lot of uncertainty. AI traffic growth is not yet well understood. There are early signals, including possible changes in uplink behaviour, but it is too soon to build 6G assumptions around one fixed traffic model. The use cases, business models and architectural implications are still evolving. This is why flexibility becomes a core 6G design principle.

NGMN’s operator view is not that everything around AI should be standardised. In fact, the opposite message came through strongly. 6G needs standardised architecture, protocols and interfaces where global scale and interoperability are required. At the same time, AI capabilities, engines and applications will need room to evolve rapidly. If 3GPP tries to standardise everything in detail, the risk is that 6G becomes too slow, too complex and too rigid for the AI era.

Another important point was migration. One lesson from 5G is that too many theoretical migration options were standardised, while only a couple of them were widely implemented in practice. For 6G, operators want a more pragmatic approach, with fewer options, clearer prioritisation and a stronger focus on real deployability. Native voice support, seamless user experience, backwards compatibility where needed and economic sustainability all need to be considered from the beginning.

Andrea Calvi from TIM then moved the discussion from 6G vision to operational reality. His presentation was based on NGMN’s work on network simplification and cloud-native, agentic AI-based operating models. NGMN’s Framework for Network Simplification is designed to help operators reduce operational complexity, lower total cost of ownership and support more sustainable network deployments.

The starting point is that network complexity has become increasingly difficult to manage. New technologies, new services and new architectures are added faster than older systems are decommissioned. The result is not just technical complexity but operational, organisational and cost complexity. This matters for 6G because a more intelligent network cannot be built on an unnecessarily complicated foundation.

Three simplification themes stood out. The first is cloud-native architecture. In the presentation, this was described as the operating system of the future network. As telecom networks become more software-based and compute-based, a common, lean and flexible cloud-native platform becomes essential. The second theme is AI-driven operations. Agentic AI needs a clean and scalable platform to run on. The third is network exposure, where operators expose capabilities through standardised APIs rather than creating a new vertical integration for every service.

This is where simplification and monetisation begin to connect. If telcos expose network capabilities in a structured way, they can become part of wider digital platforms rather than remaining confined to connectivity. This is also where edge computing may finally become more important. As AI workloads grow, it may become inefficient to send every query to a large central data centre. Intelligence may need to move closer to users, devices and enterprise sites. For operators, this creates an opportunity to combine network, edge, cloud and AI capabilities.

NGMN’s Cloud-Native Next Chapter – Agentic AI-Based Operating Models maps AI adoption levels to the CNCF Cloud Native Maturity Model and provides guidance across technology, people, skills and organisational culture. This is an important reminder that autonomous networks are not just a technology project. They also require AI-ready teams, cross-functional skills, responsible AI literacy, LLMOps processes, better observability, outcome measurement and new ways of working.

The final panel brought the themes together. Ray Le Maistre of TelecomTV moderated a discussion with Laurent Leboucher, Andrea Calvi, Kostas Chalkiotis and Itsuma Tanaka of DOCOMO Euro-Labs. The central question was what a next generation network operator might look like in the 6G era.

The answer was not a single architecture or a single business model. Instead, the discussion pointed towards a more software-driven, cloud-native, AI-enabled and platform-oriented operator. This operator still needs standards, interoperability and global scale, but it also needs to move faster than traditional telecom cycles. It needs to support continuous evolution, not just generational replacement.

Cloud-native came through as a must-have foundation. Without it, service-based architecture, modularity, scalable automation and AI-driven operations become much harder. At the same time, the panel recognised that cloud-native adoption is not yet universal and that interoperability remains a challenge. Open source initiatives such as Sylva were highlighted as important because the industry needs multi-vendor, multi-cloud and interoperable cloud platforms rather than a new generation of vertically integrated silos.

Resilience was another important theme. DOCOMO’s experience with virtualised networks, disaster recovery and self-healing capabilities showed why automation is not only about efficiency. In a more unstable world, resilience, faster recovery, energy efficiency and the ability to operate during emergencies become strategic network capabilities. Looking towards 6G, this links naturally with AI operations, sensing, satellite integration and automation in dangerous or difficult environments.

The panel also discussed fragmentation. This is one of the biggest risks for 6G. AI, cloud, satellite, network APIs, sensing and edge computing all involve ecosystems beyond traditional telecoms. Operators need to collaborate within the telecoms industry, but also across industries. Sensing is a good example. If networks are to support use cases such as drone or vehicle detection, the capability needs to work across operators, sectors and platforms. Similarly, satellite should not become a separate ecosystem that bypasses mobile operators, but should be integrated as another layer of the overall network.

The big takeaway from the NGMN track was that 6G is not just about a new radio interface. It is about how the telecoms industry adapts to AI, cloud-native operations, sustainable design, simplification, resilience and new platform opportunities. The industry has to balance two very different needs. It must preserve interoperability and global scale through standards, while also allowing faster software and AI innovation cycles.

That makes 2026 a key year. The decisions being taken now will shape whether 6G becomes another expensive technology refresh or a more flexible, sustainable and operator-relevant evolution of today’s networks. The message from NGMN’s FutureNet World track was clear: 6G must be standardised where standardisation creates value, flexible where innovation needs speed, and grounded in real operator requirements from the start.

Related Posts: 

Posted by on

Comments

Post a Comment