AI-RAN is moving from conference vision to live network results. Optus says field trials with Ericsson are improving radio performance in Australia. The work focuses on practical network gains, not distant promises.
The strongest message is simple. AI can help operators improve today’s networks without replacing major hardware. Ericsson’s software runs on existing 5G basebands and radios. That matters for operators watching capital budgets closely.
Optus highlighted three active areas during the Intelligent RAN Forum. These include dynamic link adaptation, coverage prediction, and automated coverage compensation. In plain terms, the network learns how to adjust connections in real time.
Sriharan Amirthalingam, chief technology officer for networks at Optus, explained the change clearly: “If you look at customers milling about at Circular Quay in Sydney, say, you look at their link and apply AI to optimize their experience. Which is different to what you do currently with ML algorithms – because, with AI, you adjust the antenna gains as well as modulation techniques. This is where the collaboration comes in.”
Dynamic link adaptation changes transmission rates and coding as radio conditions shift. Ericsson says this can raise throughput by up to 20 percent. It also targets spectral efficiency gains of 10 to 15 percent. In selected field cases, Ericsson reported gains reaching 25 percent.
That kind of improvement matters. Spectrum remains expensive and limited. Better spectral efficiency lets operators carry more traffic over the same assets. Users may see steadier calls, faster downloads, and fewer handover failures.
However, the story is not only about performance. Automation also reduces manual intervention. Optus described AI agents that adjust antenna settings during site outages. When the site returns, the system reverses those changes automatically.
Amirthalingam used a striking network analogy: “The network becomes a living organism – where it knows it is hurt [somewhere] and it can compensate and self-heal, and then go back to where it was before.”
Still, large-scale deployment will not be simple. SK Telecom warned that vendor differences may slow wider adoption. Each network uses different hardware, software, and data models. A successful trial may not transfer cleanly elsewhere.
Dr Dongwook Kim, director of the 6G tech team at SK Telecom, said: “Every partner has a different hardware architecture, and a different approach for AI RAN implementation. Even if one POC is successful with a specific vendor platform, it does not automatically scale across the whole network. Technologies that work well in one environment may not work the same way in another.”
That warning points to the next industry task. Operators and vendors need common test methods and reusable models. Groups such as the AI-RAN Alliance may help align use cases and validation.
For now, AI in the RAN looks most valuable when it solves current problems. Better coverage, cleaner handovers, and smarter capacity use are immediate targets. The bigger AI-native network vision can wait. Live network performance cannot.
