Creating Infinite Possibilities with Limited Wireless Resources

Huawei AgenticRAN Redefines the Value of Wireless Networks

SHANGHAI, CHINA - Media OutReach Newswire - 26 September 2025 - As global 5G-A commercialization picks up speed, Eric Zhao, Vice President and Chief Marketing Officer of Huawei's Wireless Solution, delivered a speech titled "AgenticRAN: Create Unlimited with Limited".

This was the first in-depth explanation of the AgenticRAN architecture: Based on the Three Critical Factors of "Effectiveness, Reliability, and Cost", Huawei has introduced AI into wireless networks layer by layer to generate greater network value and deliver unparalleled user experience.

Introducing AI into wireless networks layer by layer to unlock network potential

To address the dual challenges of a 100-fold increase in mobile AI traffic and green development, Huawei has proposed Three Critical Factors regarding AI adoption in wireless networks:

• Effectiveness: Focusing on the multi-scenario generalization capabilities of AI, Huawei has continuously explored and innovated in AI models and algorithms based on specific scenarios and business needs, overcoming the limitations of single-scenario AI and achieving widespread applicability across different scenarios.

• Reliability: This is to ensure long-term, stable, and trustworthy operation of AI in wireless networks, avoid unpredictable anomalies or hallucinations, and strictly control parameter setting, data privacy protection, and network security.

• Cost: Considering the need to prioritize efficiency for wireless networks, Huawei has continuously optimized the allocation of computing power and operators, introducing AI capabilities into wireless networks layer by layer to achieve the optimal balance between performance and cost.

Three directions of innovation for AgenticRAN: Using AI where it matters most

· Agentic Service: Opening up and monetizing network capabilities​
By leveraging agentic AI, Huawei has transformed traditional complex API calls into an intent-as-a-service model. This enables carriers to flexibly orchestrate multi-dimensional network capabilities using natural language, to adapt to diverse scenarios and experience requirements. Compared with conventional methods, the new model is easier to deploy and more responsive, and can significantly reduce time-to-market for new services, allowing carriers to efficiently monetize their network capabilities.

· Agentic AN: Intent-driven intelligent collaborative networks
Huawei has launched the multi-agent collaborative interface AGLink to achieve efficient coordination among multiple Executor Agents under the management of a Leader Agent. AGLink places emphasis on high reliability, high security, and privacy protection. Network optimization intentions can be directly described in natural language and then automatically translated for network deployment, significantly boosting operational efficiency. Huawei will work alongside carriers to define standards for intent interfaces and promote the standardization of A2A-T interfaces within the TM Forum (TMF) and the 3rd Generation Partnership Project (3GPP).

· Huawei adopts an adaptive approach to introducing AI and algorithms into wireless networks, continuously improving the spectral efficiency and energy efficiency of wireless communications, and flexibly utilizing time, frequency, space, and power resources to achieve the optimal balance between performance and efficiency.

Opening a new chapter in the mobile AI era

AgenticRAN will unleash the full potential of AI in scenarios where it can truly add value, delivering superior user experience and significantly increasing spectral efficiency, energy efficiency, and operational efficiency to minimize TCO. Huawei firmly believes this is just the beginning. Moving forward, Huawei will collaborate with carriers and other industry partners to transform limited wireless resources into unlimited innovation possibilities, to jointly open a new chapter for autonomous networks in the mobile AI era.

Hashtag: #Huawei

The issuer is solely responsible for the content of this announcement.

• Prev
• Next