Rakuten Mobile and T-Mobile disclosed AI-RAN automation cases that include more than 20% energy savings, 30,000 antenna adjustments during a winter storm, and Intel work to place AI inference alongside vRAN workloads.
The two operators are using different setups: Rakuten Mobile is automating network operations on its virtualized open RAN foundation, while T-Mobile is adding AI to its self-organizing network in the United States.
Rakuten Mobile’s automation stack
Rakuten Mobile says its closed-loop system has delivered more than 20% energy savings. The company also says that system has received level-four validation from TM Forum.
Rakuten Mobile is working with Intel to place AI inference alongside its vRAN workloads instead of sending everything back to a central data center.
T-Mobile’s SON and AutoPilot setup
T-Mobile’s AutoPilot solution uses intent-based automation to adjust neighboring cells when sites fail. Its Dynamic CX technology predicts demand and adjusts capacity ahead of major events.
T-Mobile says AutoPilot can make network adjustments in half the time, and its SON solution made 30,000 antenna adjustments during a January winter storm called Fern.
Read for RAN automation teams
The disclosed figures give operators concrete benchmarks for automation inside RAN control loops. They also show that both companies are tying AI to existing operational systems rather than presenting it as a standalone network layer.