Optus and Ericsson demonstrate AI-driven RAN gains in live deployment
Optus and Ericsson reported live-network improvements from AI applied to link adaptation, coverage prediction, and automated compensation, resulting in higher spectral efficiency. SK Telecom noted that scaling these AI capabilities across multi-vendor networks remains difficult. The near-term focus is on enhancing existing network performance rather than speculative AI-native infrastructure.
The live deployment in Australia targeted three specific radio functions: adjusting transmission parameters in real time based on user conditions, forecasting where coverage gaps would emerge, and automatically rebalancing signals when sites failed. Optus reported measurable spectral efficiency improvements from these AI-driven adjustments, achieved without deploying new radio hardware.
The industry distinguishes between AI embedded in networks for operational gains and networks rebuilt to serve AI traffic. Operators are prioritizing the former, viewing it as lower-risk and immediately valuable. SK Telecom's caution about scaling across multi-vendor environments underscores that interoperability standards and shared validation methods remain unresolved barriers to widespread adoption.
Mobile users in dense urban areas could experience more consistent connectivity and faster data speeds as AI-driven optimization matures. However, benefits may be uneven—operators with unified vendor ecosystems could advance quicker than those managing fragmented multi-vendor networks. The pragmatic focus on improving existing infrastructure suggests incremental consumer gains rather than transformative service changes, potentially widening the quality gap between well-resourced and constrained network operators.