Ericsson has introduced AI capabilities in its RAN, aiming to integrate AI tools directly into its Radio Access Network platforms. Unveiled on June 22, 2026, the initiative will allow Ericsson to deploy carrier-grade AI models straight into basebands and radios, enhancing 5G networks without requiring additional hardware. – Giacomo Lee/SDxCentral.
Ericsson describes its AI in RAN as a software subscription that leverages AI to enhance efficiency, performance, and energy savings. The company says the solution will strengthen 5G and 6G networks for communications service providers (CSPs) and accelerate the transition to an AI-native RAN without the need for extra hardware.
According to Ericsson, AI in RAN brings telco-grade AI models capable of real-time operation in the RAN, continuous learning powered by scalable high-quality data, and Agentic AI capabilities that enable advanced automation and network operations.
“Ericsson is redefining what’s possible in mobile networks by bringing powerful AI capabilities to service providers,” said Mårten Lerner, head of networks strategy and product management at Ericsson. With AI integrated into RAN software, we are making a significant leap toward AI-native networks, building on the AI-ready radios we introduced in February.
Ericsson highlights that its carrier-grade AI models are built for ultra-low-latency inference at the microsecond scale. “In the AI era, service providers require networks that deliver high performance, security, and efficiency,” adds Lerner. Ericsson is integrating AI into RAN to boost 5G performance and efficiency by enabling large-scale, energy-efficient AI inference.
AI in RAN integrates with Ericsson’s 22026G Advanced solution across both purpose-built and cloud-native RAN platforms, unlocking new AI-powered services. The solution is engineered to apply the most suitable AI model at the optimal location within the radio network. This is made possible through Ericsson Silicon, which delivers energy-efficient AI inference directly in the radios, combined with the latest RAN compute technology.
In addition, the portability of cloud RAN software allows AI features to be deployed seamlessly on partner platforms.
The first set of capabilities is already commercially available, having launched earlier this quarter, with additional enhancements scheduled for later in the year. These features include an AI-native scheduler for link adaptation, AI-powered macro positioning, AI-managed beamforming, AI-driven multi-layer coordination, performance management event schema files, and augmented observability for AI in RAN.
Ericsson has completed more than 15 deployments and trials globally for its AI in RAN software subscription. The company reports performance gains of up to 20 percent higher downlink throughput, up to 10 percent improved spectral efficiency, support for up to twice as many high-traffic users, 90–95 percent coverage prediction accuracy, and up to 153 times greater user-positioning precision.
The solution is already in use by several operators, including SoftBank Corp, Bell, SK Telecom, and Rogers. 17 April 2026. 09 Jun 2026.
