Skip navigation
Previous searches
    Suggested searches
      Artificial Intelligence and Radio Access Networks

      AI RAN – Smarter connectivity

      Enabling next-generation services through intelligent, adaptive networks.

      Artificial Intelligence Radio Access Networks

      Artificial Intelligence (AI) is transforming how networks are built, operated, and monetized. For Communications Service Providers (CSPs), AI is no longer optional; it is a strategic imperative. By embedding AI natively into the RAN, Ericsson enhances performance, automation and energy efficiency.

        

      AI Ran

      AI where it delivers the most value: the right model in the right place.

      Using the right model in the right place with the right compute, Ericsson’s architecture distributes intelligence across the network to maximize ROI. By matching workloads to the optimal layer, it reduces energy consumption while boosting performance. Intelligence is embedded across radios, RAN Compute, Cloud RAN within AI in RAN software and in the management layer (Service management and orchestration, SMO) with the rApps—enabling real-time optimization and energy-efficient performance across the entire network footprint.

      What is AI RAN?

      AI RAN refers to the full scope of AI integration across the RAN stack—encompassing centralized and distributed intelligence.

      Ericsson´s AI RAN strategy, combines intent-driven networking and AI-native capabilities to deliver tailored performance for AI applications to support our vision of Networks for AI, where networks enable new transformative use cases and experiences for consumer, enterprise and public sector. Gain strategic insights into scaling AI efficiently, maximizing ROI, and enabling future-proof networks.

      It includes:

      Models executed outside the RAN software (e.g., in orchestration layers or rApps) that support operations with insights, predictions, or control inputs.

      AI embedded directly within the RAN software stack (L1–L3 layers) to enable real-time, low-latency decision-making for functions like beamforming and link adaptation.

      AI on RAN enables distributed and offloaded AI processing with deterministic, low latency, high uplink connectivity, ensuring real time multimodal services for devices like smart glasses, robots, and XR applications

      Why AI RAN?

      AI RAN unlocks the full potential of networks by evolving them in two complementary dimensions:
      Networks for AI and AI for Networks. This dual transformation is key to building agile, high-performing, and future-ready networks in a virtual cycle.

      Networks for AI: By tailoring network infrastructure to the needs of AI, service providers can unlock new use cases, from immersive reality to robot assistants. This requires enhanced network capabilities, including:

      • Higher uplink capacity
      • Guaranteed quality of service (QoS) for predictable application performance
      • Consistently low latency  

      AI for Networks: By embedding AI into the network, service providers can create self-optimizing, adaptive systems that continuously improve and respond to changing demands. Trusted AI-powered software enables: 

      • Improved network performance through intelligent optimization
      • Increased automation in network operations
      • Enhanced energy efficiency
      • Faster time-to-market for new services and innovations 
      AI in RAN
      ""

      Introducing Ericsson AI in RAN

      Ericsson AI in RAN new software subscription brings the AI for Networks vision into execution by applying AI to elevate the network itself.

      • AI-native Scheduler for Link Adaptation: 20% higher downlink throughput and up to 10% improvement in spectral efficiency versus traditional rule-based link adaptation.

      • AI-managed Beamforming: 2x more high-traffic users enabled by reciprocity-based beamforming and up to 5% spectral efficiency improvement for high load.
      • AI-powered Multi-Layer Coordination: 90 - 95% accuracy identifying cell out of coverage and improves network and device resource efficiency.

      • Augmented observability for AI in RAN new functionality makes AI gains visible, measurable, and enables faster troubleshooting and confident lifecycle management of AI models with more granular, detailed and multi-dimensional data and insights.
      • Performance Management (PM) decoding schema files new functionality makes PM event data machine-readable and directly consumable by analytics, automation, and AI pipelines and AI agents. 

      • AI-powered Macro Positioning new feature unlocks new monetization opportunities such as fraud detection, emergency services, drone operations, and fleet management with 5x more accurate user location prediction.

      What’s in it for you

      Boost RAN performance and energy efficiency with AI native RAN

      AI predicts traffic patterns to improve traffic steering, to optimize spectrum efficiency and increase user throughput and reduce energy consumption.

      Accelerate your automation journey with intent-driven & AI native RAN

      AI technologies used in Ericsson solutions for automation include generative AI, digital twin, neural networks, reinforcement learning and AI agents.

      Maximize your ROI and be prepared for new subscribers and services demand

      A forward-looking AI strategy must capture today’s opportunities and remain flexible for future innovations.  AI-ready infrastructure ensures scaling AI capabilities efficiently and profitably.

      Future-proof your network with industry leading and trusted partners

      Future proof your network investments with strong ecosystem partners leading standardization and industry-wide alignment and collaboration.

      Boost RAN performance and energy efficiency with AI native RAN

      Boost RAN performance and energy efficiency with AI native RAN

      AI native RAN transforms mobile networks by embedding intelligence directly into the RAN architecture, enabling real-time optimization of performance and energy usage.

      At Ericsson, the AI journey began in the 4G era with the introduction of the first AI native features. Since then, we’ve continuously advanced mobile network performance, delivering smarter spectrum usage, improved energy efficiency, and optimized traffic steering in 5G, with ongoing innovation paving the way for 6G. Today, AI native RAN represents a major leap forward, embedding intelligence directly into the RAN to enable real-time, intent-driven optimization. 

      Ericsson’s AI native RAN software delivers measurable results for communications service providers: up to 20% improvement in throughput, 10% enhancement in spectral efficiency, and 14% energy savings through innovations like  AI Native Link Adaptation and MIMO Sleep mode. 

      Accelerate your automation journey

      Accelerate your automation journey with intent-driven & AI native RAN

      Breaking the operational complexity curve is made possible with intelligence. Distributed AI monitors and executes faster than the blink of an eye (microseconds) to optimize operations. Meanwhile, centralized AI collects long-term insights allowing for a higher degree of intelligence and further opportunities for automating and simplifying CSP’ operations and AI-powered automated solutions.

      Moving towards AI-native RAN unlocks significant advantages:

      • Scalability to handle higher levels of complexity
      • Optimized decision-making with real time decisions for more efficient network operations

      Meilke Erol-Kantarci, Strategic Product Manager, AI in RAN, Ericsson and Leonard Lee, Executive Analyst, neXt Curve

      Maximize your ROI and be prepared for new subscribers and services demand

      Ericsson’s flexible AI architecture is designed to scale with your business—delivering high performance today and ready for tomorrow’s subscriber and service demands. Intelligence at radio sites combines with centralized data centers, supporting purpose-built and Cloud RAN deployments. Open RAN interfaces give CSPs freedom of choice for seamless integration and future-proof evolution. Efficient AI scaling with minimal investment, maximum return.

      Unveil the real value of AI RAN beyond the buzz. Watch the fireside chat with Meilke Erol-Kantarci, Strategic Product Manager, AI in RAN, Ericsson and Leonard Lee, Executive Analyst, neXt Curve.

      Future-proof your network with industry leading and trusted partners

      Future-proof your network with industry leading and trusted partners

      Ericsson plays a driving role in standardization (3GPP, ETSI, O-RAN, TMF) to define the technical frameworks that ensure interoperability across vendors and operators, while also contributing to industry collaborations such as the AI-RAN Alliance  to enable energy-efficient, future-proof networks:

      • #1 RAN automation vendor in the industry ( 10+ confirmed trials in 2025)
      • 44 Members of the Ericsson Intelligent Automation Platform ecosystem
      • Largest contributor to the ORAN alliance with 18% of the technical contributions

      Ericsson is recognized in the industry for AI deployments (GSMA, Telco AI Q2 report) and is partnering with other industry players to accelerate AI developments with initiatives such as AI RAN innovation center

      Explore our solutions

      Ericsson 5G Advanced

      5G Advanced

      Ericsson 5G Advanced is a key enabler for high-performing programmable networks, which ensures consistent and superior user experience at any time and place, and influences network behavior to achieve a desired outcome.

      EIRA

      Ericsson Intelligent RAN Automation

      Our Intelligent RAN Automation is a series of technologies, solutions and services that use intelligent, machine learning to automate repetitive tasks and reduce complexity in the radio access network. 

      RAN Compute - processing platform for 5G

      RAN Compute

      Radio access network processing platform in the Ericsson Radio System, including basebands and radio processors. Ericsson latest RAN compute enhances AI models execution with more storage and processing capacity.

      Cloud RAN

      Cloud RAN

      Cloud RAN applications will enable operators to increase automation in their operation and to introduce new services to the market with speed. 

      In the spotlight

      Aim higher, Aim native

      With Ericsson AI in RAN paving te way to AI-native RAN

      Read more

      Related topical areas

      Telecom AI

      Telecom AI

      Ericsson Silicon

      Ericsson Silicon

      5G RAN

      5G RAN

      AI operations and optimization

      AI operations and optimization

      Stay updated

      News

      Blogs

      Papers

      Follow us