From the rise of ChatGPT to the impressive debut of DeepSeek, AI technology has significantly lowered its high cost barriers for training, deployment, and application through technological advancements, open-source sharing, and business model innovation. Consequently, the adoption threshold for AI has plummeted at an astonishing pace.

The rapid development of inclusive AI has presented unprecedented growth opportunities across various industries, with the telecommunications sector being no exception. Currently, operators find themselves grappling with sluggish revenue growth and operational inefficiencies that hinder high-quality network and business development. Thus, there is an urgent need to reshape network operations and business models with AI.

In the face of new AI opportunities, as the backbone of an intelligent society and the lifeblood of communication networks, the comprehensive entry of IP carrier networks into the era of intelligence has become the focal point of industry attention.

"The AI WAN solution empowers the IP carrier network in the Net5.5G era with AI through systematic innovations at three levels: AI routers, AI new connections, and AI new brains. This helps operators build networks with optimal Total Cost of Ownership (TCO), expand new businesses, enhance operational efficiency, and stimulate new business growth." On March 3, during MWC25 Barcelona, Wang Lei, President of Huawei's Data Communication Product Line, officially launched the AI WAN solution at the product and solution launch event. This move injected new momentum into the industry's intelligent transformation and garnered widespread attention.

AI Becomes a New Engine for Industry Development

AI is reshaping the telecommunications industry. Research by relevant institutions reveals that over 70% of the top 100 global operators have released AI strategies and are actively exploring the integration of networks and AI. For instance, Vodafone has introduced the "AI Framework" and is actively integrating AI technology to enhance network service quality and user experience. Similarly, Orange has set strategic goals for AI-based network operation automation to improve network operation efficiency and redefine customer experience.

The eagerness of leading operators to embrace AI stems from several reasons. Firstly, in the face of the industry-wide dilemma of "incremental growth without revenue growth," operators can leverage AI to tap into the value of massive data in the pipeline. This enables them to accurately identify business types, traffic characteristics, user behaviors, and needs, thereby creating application-level differentiated experience services and opening up new revenue streams.

Secondly, AI can significantly enhance operational efficiency and user experience for operators. For instance, AI facilitates precise network construction, ensuring that user traffic and network value are fully optimized with limited investment. By utilizing extensive network expertise, operation and maintenance experience, and vast network data, AI can proactively identify and diagnose network anomalies, automate repairs, adjustments, and optimizations, thereby significantly improving operation and maintenance efficiency and user satisfaction.

Furthermore, the rapid global development of AI has presented new opportunities for operators to expand into new businesses while also posing new challenges to WAN. The accelerated deployment of AI across various industries has given rise to new intelligent computing business requirements, such as massive sample data uploads and separated storage and computing remote training. This opens up new enterprise business markets for operators. However, AI sample data is dominated by GB-level "elephant flows" with a small number of flows, large amounts of single-flow data, and strong burstiness, which can easily cause link congestion and uneven load, reducing network throughput. Additionally, separated storage and computing remote training is highly sensitive to packet loss, necessitating the industry to leverage AI to create wide-area lossless transmission, intelligent scheduling, and optimization technologies to ensure efficient circulation of intelligent computing data and precise resource scheduling.

AI WAN Leads IP Carrier Networks into the Era of Intelligence

In the context of these challenges and opportunities, the significance of IP carrier networks as the "main transportation route" connecting mobile phone users, home broadband, and enterprise networks to intelligent computing centers and cloud services has become increasingly evident in the current AI transformation of the telecommunications industry. This is precisely why Huawei's launch of AI WAN has garnered widespread industry attention.

Wang Lei introduced that the AI WAN solution comprises a three-tier architecture: AI routers, AI new connections, and AI new brains, comprehensively enhancing network performance and intelligence levels. Among them, the AI router serves as the foundation for AI WAN's intelligent capabilities, integrating millisecond-level flow reporting, high-accuracy flow identification, and efficient security protection engines.

AI new connections provide flow-level scheduling with capabilities such as millisecond-level stable latency, ultra-high reliability, and ultra-high throughput, meeting the differentiated network needs of various applications. This helps operators provide differentiated experience services across various scenarios, including connecting people, homes, enterprises, and computing.

The AI new brain creates a network intelligent agent with a "one map and one brain" fusion architecture, assisting operators in online change simulation, fault diagnosis and handling, and improving operation and maintenance efficiency. The "one map" refers to the network digital map, which acts as a digital twin of the physical network, capable of real-time perception of extensive network data such as network topology, service traffic, and performance indicators. This realizes holographic visualization of the network and applications. The "one brain" refers to the communication large model built on the basis of the "one map." The combination of the two can quickly infer the root cause of faults and recommend optimal repair solutions, achieving an automated closed loop for operation and maintenance.

From the above three-tier architecture, it is evident that Huawei's AI WAN solution has garnered significant industry attention due to its ability to provide comprehensive AI empowerment for IP carrier networks. While AI empowerment of communication networks is not a new topic, the industry has been actively exploring and practicing it over the past few years. However, in the past, the industry mainly integrated AI technology at the component and task levels, making it difficult to form a global network insight capability, achieve full-process intelligence and automation, and flexibly adapt to increasingly diversified business scenarios. The AI WAN solution covers the network element layer, network layer, and operation layer end-to-end, realizing AI systematic empowerment of IP carrier networks. This lays a solid foundation for operators to comprehensively improve efficiency and increase revenue.

Bringing Multiple Values to Operators

Notably, AI WAN is not a future solution; it has already demonstrated its value in actual networks, aiding operators in enhancing user experience, growing business revenue, and significantly boosting operational efficiency in personal, household, and enterprise markets.

For personal services, faced with unknown base station traffic suppression and inaccurate investment expansion, AI WAN can automatically generate visual and perceptible heat maps of traffic suppression in real-time based on millisecond-level traffic collection and minute-level prediction capabilities. It also recommends precise network expansion suggestions, helping operators achieve precise investment and value-based network construction. MTN South Africa, with over 7,000 base stations, faced severe link congestion in 10% of its stations, affecting traffic experience. With the help of AI WAN's millisecond-level traffic collection and minute-level prediction capabilities, MTN obtained an efficient expansion solution. After deployment, Data Usage (DoU) surged by 25%, and traffic increased by 15.4%.

For household services, with 50% of network traffic currently encrypted, traditional signature-based matching technology has become ineffective. AI WAN's AI-driven encrypted flow identification and acceleration solution can accurately perceive applications, quickly adapt to application changes, and achieve an application recognition rate of 95%. This provides a solid guarantee for operators to create application-level deterministic and differentiated network services. Macao Telecom, in collaboration with Huawei, optimized network services through the AI computing power engine, significantly reducing game latency and enhancing customer experience. Additionally, an operator faced issues such as lag and stuttering in cloud karaoke and cloud esports services. Relying on the innovative AI WAN solution, it created a high-experience deterministic slice, providing reliable guarantees for cloud home broadband services with 1ms cloud access and second-level elastic adjustment of slice bandwidth. This not only significantly improved user experience but also attracted a large number of new cloud broadband package users, expected to significantly drive the growth of Average Revenue Per User (ARPU) values.

For enterprise services, in response to the rising frequency and scale of global DDoS attacks in recent years, with rapidly growing attack traffic and intensified intensity, traditional protection methods are struggling due to delayed detection and high miss rates. This seriously threatens the security of enterprise dedicated lines. AI WAN's intelligent flash defense utilizes millisecond-level identification algorithms and dynamic identification technology based on flow behavior libraries to improve identification accuracy to 95% and achieve second-level detection of attack flows. This can help operators provide intelligent flash defense security services to enterprise customers, with an expected revenue increase of 35%. Meanwhile, in scenarios such as massive sample data uploads and separated storage and computing remote training, Huawei's AI WAN elastic lossless network leverages AI-enabled precise traffic perception and scheduling capabilities to achieve 99% precise perception of elephant flows and balanced scheduling of millions of flows. This increases overall network throughput from 36% to 90%, effectively aiding operators in accelerating the expansion of new 2B integrated computing and networking services. Huawei and Shanghai Telecom collaborated on AI WAN solution innovation around the computing power wide-area network scenario, constructing a new end-to-end 400GE computing power wide-area network plane. This realizes efficient delivery of computing power services through key technologies such as wide-area lossless, ultimately providing computing power rental and computing power network services to industry customers.

AI is accelerating the reshaping of various industries, and this transformation is particularly urgent for the telecommunications industry. Faced with challenges such as surging connectivity demands, stagnant revenue growth, and intensifying competition, AI has become a crucial driving force for the survival and development of the telecommunications industry. The industry urgently needs to fully embrace AI-empowered networks and ecosystems to explore new revenue sources and consolidate its pivotal position as the cornerstone of an intelligent society. This is why, in recent years, the industry has never lacked the willingness for AI transformation but has struggled to find solutions that can bring significant value. From system coverage of end-to-end networks to a comprehensive focus on industry pain points and needs, and then to creating tangible value in live networks, Huawei's AI WAN solution released this time not only brings technological innovation and reflection but also injects strong momentum into the comprehensive advancement of IP carrier networks into the era of intelligence and the full implementation of AI transformation by operators. No wonder AI WAN has become the center of attention at this year's MWC.