• Coexistence of multi-RAT networks are driving spectrum inefficiencies and degraded user experience; this gets exposed more as operators transition from 4G/5G to 5G SA.

• Intelligent wireless resource management using Artificial Intelligence , machine learning and advanced computations can address these challenges.

• Huawei’s GigaBand solution is solving these with two new innovative approaches AIR Pooling (intelligent spectrum pooling and provisioning) and OptSolver, a tensor computational based multi-dimensional real-time user experience improvement.

• Huawei’s GigaBand is already commercialized successfully across the major operators globally and accelerating the transition to 5G-SA while delivering substantial economic benefits.

Efficiency and Experience are the two key performance indicators of successful cellular network. And with every data-centric cellular generation from 3G to 4G to 5G and beyond, the gamut of diverse wireless spectrum, extent of data traffic and mobile usage scenarios have grown multi-fold posing major challenges for mobile operators. To address these challenges, industry players have been contributing with noteworthy architectural as well as spectrum utilization innovations to drive the ultimate north star of exceling on efficiency and experience.

The current technological challenges for mobile operators span from serving users on multiple access technologies, fixed spectrum allocation, inefficient resource utilization and sub-optimal user experience. This slows down the migration from 4G/5G to 5G Advanced built on 3GPP Rel. 18 and Rel.19 standards. As a result, the upcoming network solutions from the ecosystem must aim to improve network performance, operational efficiency and accelerate the 5G SA ecosystem even further. 5G-A infrastructure is a key step as we march towards 6G which will proliferate over the next decade.

Intelligence in Wireless Networks is not just about Large Models

Traditionally, improving spectral efficiency has involved approaches such as increasing transmit and receive antennas (nTnR) through multi-antenna technology, which escalates Capital Expenditure (CAPEX), or waiting for 6G, which risks alienating current 4G and 5G SA users. To address these challenges, mobile equipment vendors and operators are exploring emerging AI/ML and Operation Research technologies for Air Interface (AIR) efficiency innovation.

AI deployment should be end-to-end to enhance the wireless spectrum efficiencies from RAN to terminals. The concept of wireless RAG (Retrieval Augmented Generation) becomes critical Intelligence is not just about LLMs but stems from continuous operations research and learning of network parameters for dynamically pooling, aggregating and provisioning spectrum efficiently. This involves real-time advanced mathematical computations for multi-dimension mapping of terminal’s user experience across key perform indicators.

Standardizing these AI/ML approaches is something which 3GPP intends to do in Rel.18 to Rel. 19 and beyond. However, the 3GPP is still in the nascent stages of doing this compared to some proprietary solutions already leveraging AI/ML capabilities and promising significant efficiency and user experience gains.

We recently discussed with Huawei, one of the leading network infrastructure vendors, on how they are trying to solve these challenges. It was very insightful to learn about the company’s approach with its latest GigaBand solution, which solves all of the above challenges from spectrum efficiency to user experience with multiple intelligence driven approaches. Huawei also won a GSMA Global Mobile (GLOMO) Award for this solution in March 2025.

Huawei GigaBand Brings Intelligence Driven Multidimensional Approach

This is where innovative solutions such as Huawei GigaBand aim at solving these problems by taking advanced approaches focusing on three major areas:

•    AIR Pooling Drives Intelligent Spectrum Aggregation & Utilization

This transforms the traditional AIR framework, which is typically restricted to one RAT per cell and relies on multi-cell scheduling. GigaBand introduces a multidimensional approach, intelligently combining multiple RATs within the same cell under a unified management structure. This is facilitated by a "Soft Bus for Cells Pooling" concept that uses cloud-based technologies to combine and manage logical AIRs, creating a virtual "super cell".

Resource allocation dynamically transitions from cell-level to RB (Resource Block) level, optimizing spectrum use on-demand based on real-time traffic across different RATs. This is a novel approach of using ML to transform to “service-centric” radio resource configurations. This pooling allows for centralized scheduling of multi-dimensional air interface parameters, including time-frequency resources, beamforming and transmit power. The scheduling shifts from managing spectrum and RAT resources to managing Resource Blocks (RBs).

Source: Huawei

Benefits:

• FDD+TDD in-depth convergence networking expands TDD high-frequency coverage, optimizing the Return on Investment (ROI) for 5G networks.

• Approach aims to increase spectrum utilization by 30-50%.

• Solution enables improved 5G coverage depth, for example, 3.5 GHz high-speed experience with one more wall penetration due to flexible coverage utilization.

• Up to 40% faster 5G SA uplink (UL) and 20% faster 5G SA downlink (DL) compared to fixed spectrum systems.

• Up to a 5% increase in 4G DL speed.

• Accelerates migration of LTE capacity to NR helping operator be prudent about investments.

• Provides flexible and configurable spectrum resources for emerging technologies like RedCap (lightweight 5G), passive IoT and mobile AI, facilitating large-scale deployment of innovative applications.

• Mobile OptSolver Enables On-Demand Experience Provisioning

This enhances decision-making for multi-dimension resource allocation, optimizing stable user's real-time service experience and network operational efficiency. The OptSolver utilizes advanced mathematical optimization algorithms to intelligently allocate AIR resources like time, power, spectrum and beam layers, aiming to match individual real-time user requirements and maximize resource utilization and user experience. This is powered by enhanced tensor computing hardware, which accelerates spectrum scheduling by 1,000 times compared to traditional scalar computing, enabling real-time AIR orchestration in microseconds.

Source: Huawei

Benefits:

• Multi-dimensional resource collaboration enables network planning based on peak hours and dynamic resource scheduling driven by service requirements and Service Level Agreements (SLAs).

• Reduced AI Agent Latency by 30% and Live Streaming Frozen times by 20%.

• Cell-edge mobile game latency reduced by 80% with Optsolver optimization.

Huawei's GigaBand pre-embeds 3GPP R19 technology and implements policy verification, being among the first to implement 3GPP R18 specifications and intents. R19 projects like AI/ML for Air Interface, MIMO Evolution, Network Energy Saving and Subband Full Duplex (SBFD) are relevant to GigaBand's technical features.

GigaBand Adoption & Outlook:

Huawei's GigaBand solution has been deployed or trialled in over ten commercial networks globally. Notable examples include:

• China Unicom Shandong: Pooled 700MHz and 2100MHz, resulting in a 28% improvement in 5G average downlink speed.

• China Telecom Guangdong: Deployed pooling for 1.8GHz and 2.1GHz, with pilot tests showing a 40% increase in 5G uplink speed.

• Huawei's Lab Tests (Shanghai): Pooled all Sub-3GHz bands (8 bands), enhancing uplink and downlink speeds by 40% and 20% respectively with 5G-A smartphones.

• Other commercial tests and deployments are underway in various countries including UAE, KSA, Hong Kong, Thailand, Argentina, Türkiye, and China

Outlook:

Industry leaders such as Huawei once again showcased innovative approaches to intelligence driven wireless resource management. The multidimensional approach with AIR Pooling and tensor computing based OptSolver helps operator achieve stellar spectrum efficiency and ubiquitous gigabit user experience. This will also accelerate the transition from 4G/5G to 5G SA & 5G- Advanced networks addressing the complexities of multi-RAT networks and delivering substantial economic benefits to the operators.

Huawei is one or two generations ahead of 3GPP’s efforts to bring these types of solutions into global standards and the current features covers most of the features planned in the Rel. 19 and beyond with respect to AI/ML based network management.

It is unfortunate that current geopolitics means that many operators across many regions are unable to leverage these advanced solutions from Huawei, which would help them accelerate 5G-SA.