Telecom AI

AI offers the most effective path to automating network operations and managing growing complexity from increased traffic, devices, and use cases.

Strategically deploying AI in the network has a wide range of benefits that can make significant contributions to competitiveness and business value. In broad terms, AI helps networks to deliver consistency, future-proof to support a wider range of use cases, unlock value from core and cloud technologies, achieve more efficient operations, reduce energy consumption, and maximize the performance of existing infrastructure.

Working at a scale and pace that cannot be achieved by humans alone, in networks, AI reduces the time needed to pinpoint and resolve faults. It can even predict issues before they have the chance to disrupt end-user experience.

With a strong foundation aligned to the TM Forum’s ‘level 4’ autonomous networks vision, Ericsson enables intent-driven networks powered by AI, capable of adapting in real time without human intervention. This new paradigm resolves operational conflicts and prioritizes actions based on business value, ensuring consistent performance and agility.

The 5G era sees an ongoing deepening of the integration of AI into networks. We are moving from AI as an add-on tool to AI as a foundational part of the network itself, enhancing network intelligence and establishing a robust platform for future AI applications.

AI agents will soon be deployed across the whole network ecosystem, from baseband and RAN hardware to cloud software and operations platforms. Machine learning, generative and agentic AI are driving progress towards intent-driven, autonomous networks.

In complex areas like the RAN and network operations, AI embedded directly into the radio stack and orchestration layers can now optimize critical elements like beam forming, traffic steering or MIMO sleep modes, delivering improved throughput, spectral efficiency and energy savings.

Cognitive Software meanwhile is already leveraging explainable AI and tailored AI models on cloud-native architecture to improve planning and optimization, including forecasting traffic and identifying bottlenecks to reduce CAPEX.

This journey will culminate in the 6G era with AI as a key, native network component, shaping its architecture, capabilities and services.

6G will take advantage of further advancements in hardware, compute and AI technologies, combining these to create networks that plan and anticipate with AI at their center. 6G will be fully AI-native from day one: Radical gains in energy efficiency, spectrum efficiency and uplink levels will enable transformational AI applications for consumers, enterprise and beyond.

With AI embedded at every layer, these networks will be able to learn, predict congestion, and adapt in real time to deliver tailored, guaranteed experiences. The 6G era will be defined by dynamic, autonomous networks that enable a new generation of innovators to build on top of this intelligence.

The industry target for the first commercial deployments of 6G based on ready standards is around the year 2030, by which time there is expected to be a massive uptake of mobile connectivity demanding AI usage. This is expected to apply bot to consumer AI applications like personal assistants accessed through smart XR glasses, as well as for industrial and IoT applications such as autonomous vehicles, digital twinned industrial control systems, and smart cities.

Ericsson is a pioneer in the research and development of AI in telecom, working closely with industry leaders and the wider ecosystem to shape future applications of AI technology that will improve network infrastructure, operations, energy efficiency, security, performance and customer experience.

Our next-generation AI features are already being proven in the field, with benefits ranging from 10% spectrum efficiency through AI native link adaptation, 33% energy savings from radio units through our Service Continuity AI App suite, and 25% better 5G coverage from AI embedded in RAN Compute software.

Ericsson is proactively developing infrastructure to serve external AI use cases, with R&D treating 6G as an AI-native application platform  that will see networks expose intelligence, compute and data to empower AI developers: offloading compute heavy AI tasks to the network, exposing capabilities and quality of service in AI-focused APIs, and much more.

Taking a leading role in standardization, we are not simply adding AI features on the surface , but rather, ensuring that AI applications can use networks consistently across vendors and countries, benefitting scalability and underpinning the viability of new use cases.