ZTE’s Li Xiaotong: AIR RAN Enhances Network Quality, Efficiency, and Monetization via 4xTurbo Solutions
Dual-layer AI architecture and Four Turbo engines boost spectrum, energy, and O&M efficiency while unlocking new ARPU growth through experience-driven monetization
Partner Content As global 5G commercialization enters its mature phase, operators face relentless transformation pressures: data traffic continues to surge while ARPU remains under long-term strain. "More volume, less revenue" has become an industry-wide dilemma. According to GSMA data, global 5G subscribers have surpassed 2.9 billion, with 5G data traffic now exceeding 4G. Yet simultaneously, ARPU values for mainstream operators worldwide continue to decline—the traditional "selling bandwidth" model is no longer sustainable. At this critical crossroads, Li Xiaotong, Vice President of ZTE, stated that AI is reshaping the value logic of telecom networks in a transformative way, and the leap from "connectivity pipe" to "intelligent foundation" has become the core pathway for operators to break through.
Meanwhile, the explosion of AI services is fundamentally restructuring network traffic patterns. New terminals such as AI smartphones, AI glasses, AI wearables, and embodied AI robots are emerging continuously, while various Agent applications like Doubao Phone, Qwen, Gemini Shopping, and OpenClaw are gaining traction. AI Agents are becoming the new generation of service entry points, triggering transformations in interaction models and traffic patterns. AI drives mobile networks to transition from "downlink-heavy" to "uplink-intensive", with uplink bandwidth demands jumping from 3–5 Mbps to 10–20 Mbps. Networks play a critical role in enabling real-time perception, decision-making, and feedback loops across massive numbers of terminals. Omdia predicts that by 2035, AI traffic will account for 79% of total global traffic. This means future networks will no longer be "pipes transmitting bits" but "fundamental networks delivering tokens".
Facing this historic transformation, Mr. Li emphasized ZTE's firm strategic direction of "AI & RAN deep integration". Through pioneering practice, ZTE has developed the industry-leading AIR RAN (AI-Reshaped RAN) intelligent solution. Built upon a "dual-layer intelligent architecture" that creates a closed loop of "perception-decision-execution", AIR RAN unleashes the full intelligent value of RAN through four Turbo engines. This empowers operators to enhance quality and efficiency, transforming networks into profit engines while achieving both cost reduction and revenue growth.
AIR RAN Dual-Layer Intelligent Architecture: Truly Integrating AI into Wireless Networks
Mr. Li introduced ZTE's industry-first introduction of the intelligent computing engine AIREngine within base stations, pioneering a "site & Network" dual-layer intelligent collaborative architecture. On the base station side, AIREngine enables millisecond-level service perception and local decision-making. It can real-time identify the service types, terminal status, and channel quality—outputting precise KQI and root cause analysis data to provide upper-layer intelligent agents with "clear visibility and accurate judgment" raw data. On the network side, AI Agents are built upon Large Language Models (LLM), including Experience Agent, Fault Agent, Optimization Agent, and Energy Agent. These agents leverage knowledge graphs, chain-of-thought (CoT) reasoning, and digital twin technologies to possess "human-like" decision-making capabilities for intent understanding, policy generation, capability scheduling, and closed-loop evaluation.
This architecture is compatible with existing networks and plug-and-play, enabling low-cost, high-efficiency intelligent upgrades while providing operators with an AI evolution path that is deployable, monetizable, and sustainable.
Four Turbo Engines: From Cost Reduction to Value Creation
Based on the dual-layer intelligent architecture, ZTE's AIR RAN innovatively introduces four Turbo solutions, comprehensively covering the three major network efficiencies (spectral efficiency, energy efficiency, and O&M efficiency) as well as user experience and commercial monetization—truly achieving "technology empowering business."
SpectrumTurbo: 20% Spectral Efficiency Improvement, Maximizing Value Per Hertz
Facing growing spectrum constraints, SpectrumTurbo leverages AI-driven dynamic smart cell adaptation to evolve from 4-cell to 6-cell splitting per sector in FDD M-MIMO, enabling a software-only upgrade that boosts cell capacity by over 15%. TDD M-MIMO solution leverages AI-enhanced SRS channel estimation, traffic prediction, and resource orchestration to achieve a 20% improvement in 4G/5G spectral efficiency. Mr. Li noted that this solution, already deployed across multiple global networks, enables operators to unlock network potential without deploying additional sites or acquiring new spectrum.
EnergyTurbo: Extra 15% Power Consumption Reduction, Achieving Optimal Efficiency at any Load
EnergyTurbo is an AI-driven energy-saving solution based on a dual-layer intelligent architecture, integrating AIREngine and Energy Agent to create a closed loop from perception and decision-making to execution, achieving optimal energy efficiency and simplified operations across all scenarios. Building on precise traffic prediction, parameter self-optimization, and rapid evaluation assurance, AIREngine enables second-level traffic prediction and precise service identification, delivering differentiated energy-saving strategies for diverse services to ensure both energy savings and user experience. The Energy Agent enhances operational efficiency and further improves energy-saving gains through intent-based policy orchestration and digital twin simulation. In the Thailand project, overall energy consumption was reduced by 14.57%, saving 2.25 million kWh annually, achieving a win-win outcome for sustainability and economics.
OpsTurbo: 20% MTTR Reduction, 20% Zero-touch Ticket Resolution Rate, Entering the "Agent-Led, Human-Assisted" New Era
OpsTurbo encompasses two sub-solutions: FM (Fault Management) and PM (Performance Management).
In FM scenarios, a "large-small model collaboration" architecture is adopted: small models rapidly process massive real-time data to complete alarm correlation and root cause analysis; large models (Fault Expert) combine knowledge graphs and CoT for complex reasoning, pushing diagnostic results and operation guidelines to engineer mobile apps to achieve "front-back-end integration", reducing MTTR by 20%. Mr. Li introduced that this solution helped China Mobile achieve TM Forum certification with an L3.8 high score in RAN fault management—the highest in the industry.
In PM scenarios, the network optimization expert Agent enables end-to-end automation of quality analysis, solution generation, and closed-loop execution, with engineers only required for key node confirmations, achieving a 20% zero-touch ticket resolution rate.This helped China Mobile achieve TM Forum certification with an L3.95 score in RAN network optimization—also the highest in the industry.
ExperienceTurbo: 5%-20% ARPU Improvement, From "Selling Traffic" to "Selling Experience"
ExperienceTurbo is the core enabler for operators' value transformation, Mr. Li emphasized. Through AI's precise identification of high-value services (such as live streaming and gaming), it achieves full-process closed-loop "pre-subscription assessment, in-process dynamic assurance, and post-experience evaluation," driving 5G/5G-A package upgrades. 5G-A new packages have been widely deployed in China, covering scenario-based add-on packages for concerts, sports events, campuses. At a Thailand music festival, ExperienceTurbo significantly improved user experiences for TikTok and Line applications, reducing buffering and increasing throughput. Mr. Li stated that this represents not only a technological breakthrough but also a business model innovation—operators shifting from "traffic billing" to "experience subscription," bringing 5%-20% ARPU value improvement. And based on technologies such as L4S and latency-prioritized scheduling, ZTE has completed the first commercial verification of XR immersive services, achieving a stable 90% reduction in latency and a significant decrease in service response time under congested scenarios, accelerating the commercial deployment of 5G-A networks for emerging AI services such as XR, cloud gaming, and AI glasses.
From Network Modernization to Network Monetization: AIR RAN Defines the New Paradigm
The global communications industry is transitioning from the "network construction era" to the "intelligent monetization era". ZTE's AIR RAN, with "dual-layer intelligence" as its architecture and "Four Turbos" as its traction, not only solves operators' current cost reduction and efficiency challenges but also opens new revenue growth curves for the AI era.
Mr. Li concluded: "We firmly believe that future networks will not be 'wider pipes' but 'smarter partners'. ZTE is ready to join hands with global operators, using AIR RAN as the engine, departing from network modernization and advancing toward network monetization, moving toward the intelligent future of 5G-A and 6G—making every watt create value, every connection carry intelligence, and every investment yield returns."
The future has arrived; the intelligent journey begins. ZTE AIR RAN empowers network quality enhancement and efficiency improvement, enabling networks to become a source of profit.
Contributed by ZTE.