Nokia and Amazon Web Services are piloting AI agents that adjust 5G network slices in real time. The tests mark a shift from manual configuration toward automated operational control in live telecom environments.
According to a joint announcement from Nokia, the system is being tested with du in the United Arab Emirates and Orange across Europe and Africa. The setup combines Nokia’s network slicing and automation tools with AI models delivered through Amazon Bedrock, AWS’s managed AI service platform. The companies describe the architecture as “agentic AI.”
Can AI Agents Monetize 5G Network Slicing?
Network slicing allows multiple virtual networks to run on shared physical infrastructure, each optimized for distinct use cases such as emergency response or high-bandwidth consumer traffic. While slicing is embedded in the 5G standard, operators have often relied on fixed configurations and manual planning. The pilot introduces AI agents that monitor latency, congestion, event schedules, and weather data to adjust network resources automatically.
Research firm GSMA Intelligence has noted that many operators view slicing as a potential enterprise revenue driver, yet adoption has lagged due to operational complexity and uncertain demand. If AI systems can dynamically reallocate capacity during spikes such as stadium events or disaster response, operators may be able to offer temporary service-level guarantees without manual intervention. Would automated scaling make enterprise 5G behave more like cloud infrastructure?
Orange has previously stated that enterprise customers expect connectivity to scale on demand, similar to cloud computing. The involvement of AWS reflects a broader trend of telecom workloads shifting toward public cloud environments. Analysts at Dell’Oro Group report rising telecom cloud spending as operators modernize core networks and adopt software-driven architectures.
Still, the technology remains in pilot phases, with demonstrations and limited rollouts rather than full production deployments. Regulators and operators will likely scrutinize how AI-driven control loops are supervised, particularly given telecom networks’ role in critical communications. The next milestone will be whether pilots convert into commercial contracts with defined service-level commitments.
