AI-RAN Alliance shifts focus to deployment plans and edge revenue models
The AI-RAN Alliance is moving from demonstrations toward implementation blueprints, test frameworks, and commercial models for AI in mobile networks. Its working group on AI on RAN now covers edge inference, physical AI, semantic communications, and multimodal RF sensing and ISAC. The group is also examining security, AI workload placement, and potential operator revenue from edge compute, inference, sensing, and latency-guaranteed connectivity.
The group is led by Alex Jinsung Choi from SoftBank’s technology unit and CTO office. It includes more than 142 member organizations across over 17 countries. Its efforts are divided into three areas: improving radio performance with AI, sharing infrastructure between radio and AI computing, and delivering AI services through networks and edge infrastructure.
Its AI-on-RAN work is structured around services rather than technologies alone. It considers the latency, privacy, and compute needs of each service. Decisions about where to run AI must weigh model size, data location, transport costs, and available resources. External AI workloads also require separation from core radio functions.
If operators turn mobile networks into distributed AI platforms, people and businesses near edge sites could see faster, more localized inference, sensing, and latency-sensitive applications. Developers and enterprises may gain new deployment options, while operators could diversify revenue beyond connectivity. At the same time, broader use of edge AI and RF sensing may raise privacy, security, and access questions. Actual outcomes may depend on demand, computing capacity, reliable service delivery, and how carefully third-party workloads are isolated from critical network functions.