Level 4 Autonomous RAN – From Vision to Early Reality

The analyst blog argues that growing RAN complexity and constrained revenue growth are pushing operators toward intelligent automation, including AI-driven steps toward Level 4 autonomy, with early deployments in targeted domains and validation tied to energy efficiency.

Market and operating context

The note says mobile networks are expanding their bandwidth needs, frequency/technology diversity, and use cases spanning eMBB, FWA, private wireless, and IoT, increasing RAN operational complexity. It also reports that revenue growth has constrained operators’ ability to scale operating expenses alongside the complexity.

The blog cites Ericsson Mobility Report data that mobile data traffic rose 16% in 2025, while Dell’Oro Capex Report figures show operator revenue growth at 1% CAGR over the past ten years.

How the TM Forum autonomy framework is used

The report frames RAN autonomy using the TM Forum scale from L0 to L5, with humans operating at L0 and full AI/automation with no human intervention targeted at L5. It describes progression as moving from task automation toward automating decisions, with a reference to operators and vendors aligning on the transition.

The blog defines intermediate levels as ranging from manual and assisted operations at L1, to partial automation at L2, to conditional closed-loop operations at L3, to high autonomy with near-full automation for many scenarios at L4.

Automation phase progression and expectations

The note describes an earlier evolution from 4G and early 5G automation focused on provisioning, configuration management, software upgrades, and alarm handling. It then says 5G deployments expanded automation into optimization use cases such as mobility management, load balancing, energy efficiency, and performance tuning using policy engines and domain analytics.

It characterizes the next phase as AI-driven autonomy, where systems adapt dynamically rather than rely only on predefined rules, and links Level 4 to “zero-touch networks” that detect and fix issues before they become major problems.

Reported benefits and supplier performance claims

The blog lists RAN automation goals that include maximizing ROI on network investment, improving performance and experience, boosting network quality, reducing complexity, and reducing energy use. It also references CO2 emissions reduction as an operator goal.

It compiles performance claims from vendors and operators, including Ericsson estimating 15% improved spectral efficiency from Intelligent RAN Automation, Huawei demonstrating “up to 50%” user experience improvements in some settings, and Nokia reporting “up to 80% efficiency gains with zero-touch radio network optimization.”

The note also cites ZTE demonstrations for large-scale L2/L3 automation, including “around a 30% reduction in fault recovery times” and “roughly 20%+ improvements in resource utilization,” and says Samsung incorporated AI and automation into its 5G RAN portfolio while large-scale L4 deployments were more limited than some larger suppliers.

Level 4 status in live networks

The blog states that most of the industry remains at TM Forum Level 1–2, where automation is largely domain-specific and rule-based rather than full network-wide autonomy. It says Level 4 is “no longer theoretical,” with implementations limited to a small group of leading operators.

It cites Rakuten Mobile as having demonstrated Level 4 at scale in a live RAN network and says TM Forum validation for RAN energy efficiency found Rakuten able to realize “20% RAN energy savings using AI-driven closed-loop control with no impact on customer experience.”

Named certifications, assessments, and deployment examples

In June 2025, the blog says TDC NET and Ericsson achieved “TM Forum Level 4 autonomy certifications for a live RAN deployment,” and that validation focused on Ericsson’s PCEM software, which reduced the energy required to transmit 1 GB of data by “approximately 5% under live network conditions.”

The note reports China Mobile used TM Forum ANLAV assessments at Innovate Asia 2025 and received “TM Forum certifications across multiple use cases” including service assurance, wireless energy optimization, and IP fault management. It also describes China Telecom, China Unicom, and Huawei as pursuing Level 4 in optimization, assurance, and energy savings use cases, with Huawei citing “10+ international operators” implementing L4 autonomous networking in live production environments.

It further says Nokia and STC demonstrated Level 4 autonomy in live RAN operations, and during the Hajj period with traffic increasing by 40% the network executed “10K autonomous operations per hour,” helping improve DL throughput by “approximately 10%.”

Roadmap from single-domain to cross-domain autonomy

The report says most Level 4 implementations are scenario-based rather than network-wide, and it divides the path into Phase 1 near-term single-domain scenarios and Phase 2 medium-term cross-domain autonomy. Phase 1 includes RAN closed-loop automation, AI-assisted decision-making with human-defined guardrails, limited cross-domain coordination, and energy optimization plus automated fault detection and traffic/load optimization.

Phase 2 is described as extending autonomy across domains to enable end-to-end service orchestration, including multi-domain coordination across RAN, transport, and core, intent-driven automation, minimal human intervention, and end-to-end assurance with cross-domain root-cause analysis and mobile network optimization.

Role of RAN agents

The note describes RAN agents as an enabler for moving from rule-based automation to a goal-driven model, saying agents “can interpret intent, assess real-time conditions, and take action while continuously learning from new data.” It adds that the aim is to enable context-aware decision-making and real-time adaptation rather than static workflows.

It then states that near-term RAN agents enhance domain-level automation through scalable operations, and that the “real step change” is expected from cross-domain collaboration where distributed agents coordinate across RAN, transport, and core. The blog also says this multi-agent approach introduces “new challenges around coordination, interoperability, and building trust in autonomous decision-making.”

Analyst outlook and adoption constraints

The blog concludes that the shift toward RAN automation is occurring, but it characterizes the journey as taking longer than some operators promised for large-scale Level 4 by 2025. It says autonomy today is achieved within well-defined guardrails with humans setting policies and objectives.

The note frames adoption pace as uncertain and says the transition to Level 4 is a human challenge as well as a technical one, alongside limited revenue upside. It states that RAN automation will play a growing role in the second half of 5G and likely from the start with 6G.

This Analyst Signals brief reflects a neutral, fact-based summary of the original research note.