“The industry is moving beyond generative AI experimentation toward systems capable of taking operational action.”

This Research Note summarizes Appledore’s insights from showfloor and after-hours discussions at Mobile World Congress 2026, where the dominant tone was execution-focused: moving from GenAI experimentation to Agentic AI that takes real operational actions. Conversations centered on data sovereignty, AI factories, edge inferencing in the RAN, agentic AI for network automation, and how CSPs can monetize AI. 6G and quantum were present but muted; the near-term spotlight was on software-driven autonomy, data fabrics, and decisioning that improves network efficiency and reliability.

BSS and digital enablement discussions revolved around agents, data/ontology layers, and outcome-based pricing. Practical challenges include siloed data, unclear agent definitions, and measurement of AI-driven outcomes. Additional themes included AI-on-RAN business cases, AI-native networks as an evolution of 5G-Advanced, security/SNOC convergence and post-quantum operationalization, digital twins at the core of autonomy, Open RAN’s single-vendor reality, OCUDU’s open framework push, slow-burn network API monetization, and a shift from bespoke enterprise projects to engineered-for-scale offers.

Companies referenced in this report include: Huawei, Amdocs, Google, Google Cloud, Netcracker, Deutsche Telekom, Vodafone, Microsoft, Nokia, Ericsson, T-Mobile, Telenor, MTN, Ooredoo, Globe, Viavi, Wavelo, AWS, Intel, ENEA, Mercedes-Benz, Siemens, AMD, NVIDIA, AT&T, SoftBank Corp., Cisco, Red Hat, Telefonica, Linux Foundation.

You can also listen in as the team provide a hour-long debrief from MWC26 on our Youtube channel, Talking Transformation.