Nokia and Amazon Web Services (AWS) have jointly introduced what they describe as the first agentic AI-powered 5G-Advanced network slicing solution deployed in a live 5G network context. The development was showcased at the Mobile World Congress (MWC) 2026 event in Barcelona.
The trial is already underway with du and Orange, who are putting the system through its paces on their own networks.
The initiative combines Nokia’s network slicing and 5G technology stack with AWS’s AI platform and cloud infrastructure to deliver an intent-based 5G network slicing capability powered by agentic artificial intelligence. Operators du and Orange are among the first to trial the solution on live networks.
According to official descriptions, the agentic AI system ingests external contextual inputs, such as geographic data, traffic patterns, event schedules, and network key performance indicators, to infer optimal configuration policies for slicing the network. This system then automates adjustments across radio access and transport layers to meet service-specific requirements.
Pallavi Mahajan, Nokia’s Chief Technology and AI Officer, said, “By fusing our advanced network slicing with agentic AI, operators can finally deliver intent-based, premium services that intuitively align with real-world demands. It’s about unlocking new value for telcos and opening the door to the next wave of enterprise, industrial, and consumer applications.”
Amir Rao, AWS’s Global Director, GTM & Telco Solutions, said, “Until now, manual tweaks and rigid policies stood in the way of true on-demand service. By harnessing agentic AI via Amazon Bedrock within Nokia’s platform, we’re moving past technical hype into genuine business enablement, letting operators finally monetize 5G investments through services that adapt, automatically and in real time, to whatever the world throws at them. We see this as the dawn of real-time, intent-driven service for telecom customers.”
Network slicing: the logical partitioning of mobile network resources for distinct use cases, has been a strategic objective of 5G since its inception.
The announcement arrives amid broader industry trends toward AI-assisted network operations where machine learning and inference engines augment human network planners. However, real-world deployments of agentic AI mechanisms remain limited. This Nokia-AWS collaboration appears to be among the first public demonstrations of autonomous slice orchestration in a commercial telco context.
Industry sources suggest that most current slicing implementations still depend heavily on static or semi-automated policies, with limited integration of external contextual data. The innovation here lies in real-time decision loops where external signals, such as geographic or user behaviour data, can trigger adjustments in network segmentation and resource allocation.
In statements tied to the demonstrations, technical executives from both du and Orange emphasised the potential of intent-based slicing to support enterprise and consumer use cases with improved responsiveness.
Saleem Alblooshi, du’s Chief Technology Officer, said, “We’re proud to be among the first to put this technology through its paces. Agentic AI-powered slicing means we can deliver not only mission-critical enterprise connectivity, but also richer experiences for everyday consumers, responsive, premium, and reliable.”
Atoosa Hatefi, Orange’s Director of Innovation in Radio and Environment, added, “Orange has always been at the cutting edge of 5G innovation. This trial shows how AI-driven slicing lets us see around corners, anticipating needs and delivering services tailored not just to industries, but to immersive entertainment and beyond.”
Telecommunications infrastructure is increasingly positioned to incorporate AI beyond conventional automation tools. Autonomous network functions, including slice orchestration, fault management, and capacity planning, are viewed as enablers for future 5G-Advanced and eventual 6G networks.
Conclusion
While the technological innovation is significant, its operational integration and commercial impact will depend on further trials, standardisation progress, and tangible business models over the coming years.
