Evolution from 5G to 6G: Key Technologies for Future Networks

Commercial 5G networks are transitioning from basic coverage to deepening enablement, and large-scale 5G-A deployment has laid a connectivity foundation for industry digital transformation. The explosion of AI applications, the rise of new services such as the low-altitude economy and integrated communication and sensing, and the urgent need for air-space-ground coverage are driving mobile networks from "Internet of Everything" to "Intelligent Convergence." To meet emerging requirements such as massive uplink, deterministic connectivity, compute-network convergence, and ubiquitous wide-area connectivity, wireless networks require systematic upgrades across architecture, air interface, and computing power. This paper explores the evolution of wireless sites toward 6G, with a focus on key technologies including massive MIMO, integrated sensing and communication (ISAC), cellular passive IoT, subband full duplex, AI-RAN collaboration, and satellite phased arrays.

Key Technologies for Evolution from 5G to 6G

• Massive MIMO Antenna Array

Massive MIMO, one of the most representative 5G technologies, has achieved widespread commercial deployment, with mainstream configurations of 32TR and 64TR, typically comprising 192 and 384 antenna elements, respectively.

The U6G (6425-7125 MHz) band is a critical frequency band for 6G. With high propagation loss and short wavelengths, it accommodates up to four times as many array elements as the 3.5 GHz band within the same physical area, easily exceeding one thousand elements. In 2026, ZTE unveiled the industry's first 1500+ element U6G ultra-massive MIMO solution at MWC, adopting a hybrid analog-digital beamforming architecture that balances beamforming accuracy with RF channel costs, providing a practical path toward U6G coverage comparable to 3.5 GHz while enabling a two-fold performance improvement. The current U6G industry chain is maturing, and related products are ready for early commercial deployment in 5G-A networks.

• ISAC

5G-A ISAC integrates communication and sensing within a single base station, utilizing enhanced air interfaces to enable target detection and tracking. Currently undergoing pilot projects in multiple locations across China, it is driving emerging applications such as low-altitude security, low-altitude economy, waterway navigation, and maritime port supervision, evolving from technical validation to large-scale application and laying the foundation for 6G all-domain sensing and intelligent networks.

For 6G ISAC, we are committed to building an integrated system encompassing communication, sensing, positioning, and multimodal fusion. The network is centered on base stations to achieve all-domain communication and advanced sensing. It integrates multimodal data from sources such as optoelectronics and spectrum detection, while leveraging AI to enhance sensing performance. Through high-precision environmental modeling and ray tracing, 6G networks can construct digital twin platforms to enable real-time mapping between physical and virtual spaces.

In the field of 5G-A ISAC, ZTE has pioneered a hybrid waveform combining continuous waves and pulsed waves, launched a self-transmitting/self-recieving wide-angle product to address seamless low-altitude sensing challenges, and continues to explore multimodal fusion solutions based on end-edge-cloud collaboration.

• Cellular Passive Internet of Things (A-IoT)

Traditional NB-IoT and RedCap rely on active devices and are unsuitable for passive IoT scenarios. RFID primarily supports point-to-point communication, with limitations in coverage, inventory efficiency, and accuracy that make it difficult to meet the requirements of the Internet of Everything. Cellular passive IoT integrates cellular and passive technologies, combining advantages such as continuous coverage, massive connectivity, low cost, low power consumption, and maintenance-free operation.

Cellular passive IoT base stations enhance the read-write range through interference cancellation, flexible energy storage, and new air-interface coding. They support ultra-massive concurrent tag access and improve inventory efficiency by leveraging time division multiple access, frequency division multiple access, multi-rate transmission, anti-collision scheduling, and group-based inventory while enabling value-added functions such as sensing and positioning. Their low cost, low power consumption, and massive connectivity provide a data foundation for 6G applications such as digital twins, the metaverse, big data, and artificial intelligence.

• Subband Full Duplex and Dynamic Frame Structure

AI is reshaping the terminal ecosystem, with AI smartphones, AI glasses, smart vehicles, and embodied intelligence continuously emerging. Mobile communications are evolving from "connecting people" to "connecting intelligent devices," creating new requirements for deterministic connectivity, high uplink capacity, and computing-network convergence. To address these challenges, ZTE has introduced two innovative technologies for the 5G-A phase:

•    Subband full duplex (SBFD): Divides a single carrier into multiple subbands, with each subband independently configurable for time slots, enabling flexible uplink and downlink resource scheduling. ZTE has completed the R&D and deployment validation of pre-commercial SBFD small cells, achieving over 60% latency reduction, more than 4 dB uplink coverage improvement, and over 70% uplink throughput increase. This technology is being extended to macro site scenarios to support broader network deployments.
•     Dynamic frame structure: Based on AI-predicted service load, the frame structure is dynamically adjusted to align resource allocation with service requirements. By leveraging diverse anti-interference technologies to address coexistence issues between heterogeneous frames, ZTE and operators have successfully completed pre-commercial trials in macro-micro hybrid scenarios, and the solutions have received multiple industry recognitions.

The two innovations jointly leverage time and frequency domains to meet the requirements of uplink-heavy scenarios, laying a foundation for 5G-A deployment and 6G evolution.

• AI and RAN Convergence

The 5G-A phase has initially realized AI-enhanced RAN: AI use cases such as L1 CSI compression feedback and beam management have been standardized, while intelligent functions including L2/L3 intelligent link adaptation, intelligent energy saving, and automated operations and maintenance have been commercially deployed. However, current external AI solutions face pain points such as dispersed computing power and limited protocol adaptation. 6G will evolve towards AI-native RAN by embedding AI into the protocol stack to improve spectrum efficiency, user experience, and energy efficiency, thereby empowering various AI applications.

ZTE has launched its AIR RAN intelligent solution, which leverages a heterogeneous hardware architecture that integrates general-purpose and dedicated computing resources to balance computing power and overhead. Its self-developed intelligent computing boards have been deployed in commercial scenarios such as energy saving, operations and maintenance, and experience optimization. Meanwhile, ZTE is advancing research on 6G AI-native architectures, developing proprietary AI acceleration chips, and building an integrated communication-computing-intelligence system capable of smooth upgrades.

• NTN Phased Array Antenna

Based on 3GPP R17/R18, 5G NTN has entered the commercial deployment phase. As a key hardware component, phased array antennas achieve rapid beam switching through coordinated phase control among array elements, overcoming the limitations of mechanical scanning. A single satellite can generate over 100 independent beams, filling coverage gaps and enhancing capacity in high-density areas, with key applications in maritime, aviation, and emergency communication.

To address challenges such as limited access capacity, high costs, and significant latency during inter-satellite handovers, phased array antenna technology is evolving rapidly. Ultra-large-scale multi-beam antenna arrays support massive access; the integration of industrial-grade silicon-based chips with high-gain antennas drives miniaturization, lightweighting, and cost reduction; and multi-satellite collaborative scheduling achieves seamless, low-latency, all-domain access.

As a key driver of global NTN development, ZTE has collaborated with partners across the chip, module, and application sectors to complete end-to-end technical validation, build a comprehensive NTN industrial ecosystem, and develop proprietary phased array antennas covering S-, Ku-, and Ka-bands, accelerating the evolution toward integrated space-ground networks.

Future Evolution Trends of Wireless Sites

Based on the above six technical directions, wireless sites are evolving from 5G "high-performance connectivity nodes" to 6G "intelligent convergence hubs," which can be summarized into four key trends.

• Architectural Trend: From Base Stations to Integrated "Connectivity, Sensing, Computing, and Intelligence" Platforms

Traditional base stations focus on communication connectivity, whereas in the 6G era, they will evolve into integrated platforms combining communication, sensing, computing, and intelligence, transforming from bit pipes to edge intelligent nodes capable of environmental perception, data computation, and intelligent decision-making. Future wireless sites will be centered on base stations, enabling efficient all-domain communication and advanced sensing, deeply integrating multimodal sensing data, and building a service-oriented RAN architecture.

• Air Interface Trend: From Static Frame Structures to Dynamic Intelligent Air Interfaces

The 5G frame structure is relatively fixed with static resource allocation, whereas the 6G air interface is evolving towards a dynamic and intelligent paradigm. Subband full duplex and dynamic frame structure serve as two enablers in the time-frequency domain to flexibly address diverse requirements such as high-capacity uplink, low latency, and deterministic services. 3GPP Release 20 has designated AI-native air interface design as a 6G study item. The future air interface will evolve towards a closed loop of "real-time sensing, intelligent decision-making, and dynamic adaptation," with lightweight AI models embedded across the protocol stack to fundamentally address challenges in energy efficiency and spectrum efficiency.

• Coverage Trend: From Ground-Based to All-Domain Coverage Across Air, Space, Land, and Sea

5G relies primarily on terrestrial base stations, with coverage constrained by ground-based infrastructure. 6G will enable integrated coverage across air, space, land, and sea. NTN phased array antennas enable direct satellite-to-mobile connectivity, while ultra-large-scale antenna arrays with thousands of elements enhance high-frequency coverage, enabling U6G coverage performance comparable to 3.5 GHz. The future coverage boundary will extend from "ground" to "sky," expand from "urban areas" to "oceans and deserts," and achieve seamless access across all domains.

• Computing Trend: From Dedicated Hardware to Heterogeneous Computing Platforms

5G base station computing resources are primarily dedicated to baseband processing, relying mainly on dedicated ASICs. For 6G, wireless sites are evolving into distributed edge computing nodes. The deep integration of AI and RAN requires base stations to support AI inference while ensuring deterministic communication. Dedicated ASICs ensure real-time communication performance and energy efficiency, AI acceleration chips provide native AI capabilities, and general-purpose computing supports flexible deployment of edge AI services. Heterogeneous computing architectures enable wireless sites to achieve efficient communication while supporting diverse workloads such as AI inference, sensing, and edge computing.

Conclusion: The Intelligent Transformation of Wireless Sites

The transition from 5G to 6G represents a paradigm shift from "Internet of Everything" to "Intelligent Integration."

Commercial 5G networks have established a solid foundation for industrial digital transformation. Meanwhile, breakthroughs and validations in key technologies such as Massive MIMO, integrated sensing and communication, cellular passive IoT, SBFD, AI-RAN convergence, and NTN phased arrays are laying the technical foundation for 6G. In 2026, 3GPP officially approved the timeline for the first 6G specifications, Release 21, marking the transition from conceptual exploration to system verification and standardization. The first commercial services are expected at the end of 2029 or early 2030.

In response to this transformation, wireless site technology upgrades should take 5G-Advanced as a key evolutionary step to facilitate ecosystem integration and lay the technical groundwork. Through continuous breakthroughs in areas such as ultra-massive MIMO, integrated sensing and communication, AI-native networks, and space-air-ground integration, together with scenario-driven innovation and technology-driven applications, we can build an intelligent, ubiquitous, and continuously evolving 6G network, providing a solid connectivity foundation and computing support for the intelligent society in the 2030s.