The 2025 China AIGC Innovation Application Forum of the China International Fair for Trade in Services was held in Beijing on September 12, 2025, themed "From Large Models to Intelligent Agents, Driving AI New Ecosystem." Wu Riping, Network Product Solutions Manager for Ericsson China, attended and delivered a speech.

The following is a transcript of the speech:

Good morning, leaders and experts. I am delighted to represent Ericsson at this year's trade fair. For those unfamiliar with Ericsson, it belongs to the telecommunications industry. Founded in 1876, Ericsson has become one of the world's leading telecommunications equipment service providers and a global setter of communication standards. Today I want to report on the theme "When Communications Meets AI" - the transformation of autonomous intelligent networks in mobile communications in the AI era.

From the telecommunications industry perspective, we have observed milestone events and figures over the past few years. The Mobile World Congress is the industry's biggest event, and AI has been the most frequently mentioned buzzword over the past two years. More specifically, through interviews with 100 leading global operators, 65% have already formulated clear AI transformation strategies and begun implementation.

For large operators, this proportion rises to 88%, and we see cross-industry alliances and cooperation developing rapidly. Notable is the AI-RAN alliance jointly participated by NVIDIA and Ericsson, with such alliances increasingly collaborating. Standardization is crucial in the telecommunications industry, and AI has become one of six major application scenarios in overall network evolution.

From industry trends, based on the TMF protocol architecture - TMF being the autonomous network standard setter - by September this year, over 76 leading industry collaboration players have announced support for the telecommunications industry's AI transformation and autonomous network development.

The combination of telecommunications and AI isn't just because AI is trendy, but because we see strong inherent driving forces, manifesting in two aspects:

First, AI empowering networks - the transformation from traditional networks to what we call autonomous intelligent networks. Benefits include efficiency improvements and new business services generation. Key technologies include large language models, intelligent agents, and widespread use of digital natives in the telecommunications industry, gaining broad industry consensus and rapid development.

Second, equally exciting is networks empowering AI. Our vision is leveraging high-performance mobile communication networks to enable increasingly widespread mobile AI applications. Core technologies include differentiated connectivity and cloud-edge-device computing power coordination architectures. This area is in the demand exploration stage but shows tremendous growth potential.

Today I'll focus on the first part - AI empowering networks, or autonomous intelligent networks. These develop through five stages, somewhat like autonomous driving levels L0 to L5, from completely manual (L0) to fully autonomous (L5).

The key milestone is the L3-L4 transition. L1-L3 involves system-assisted manual operations to improve efficiency, while L4 and beyond features autonomous network operations under human supervision - a higher-dimensional achievement. Effects are evident: L1-L3 mainly saves network costs, while L4 brings new business launches and generation.

Currently, globally, major operators are mainly between L1-L3, with only 4% of operator scenarios reaching L4. However, looking to 2030, over 80% of operator scenarios are expected to reach L4. The telecommunications industry being the world's largest foundational infrastructure represents enormous potential and challenge for AI deployment.

As Ericsson, we're pleased that as a telecommunications industry leader, we've made excellent progress in autonomous intelligent networks. We're doing four things:

First, products - Ericsson's new generation communication products have fully achieved AI enablement, creating excellent autonomous network architecture and foundation.

Second, standards and ecosystems are crucial in telecommunications. Success comes from global collaboration and open ecosystems, not individual company standards. This applies to autonomous networks, where Ericsson contributes extensively to standards and ecosystem development.

Third, telecommunications is highly professional with extremely high reliability requirements. Network failures can instantly become headlines, affecting all businesses. Combining telecommunications and AI requires both AI capabilities and professional telecommunications expertise. Ericsson is building and already possesses strong integrated communications and AI capabilities, with autonomous networks being a key R&D investment focus over the past five years and continuing forward.

What specific examples do we have? We're pleased to share that after years of effort, Ericsson has moved autonomous intelligent networks from concept to commercial reality - a highly challenging achievement. Autonomous networks from L0-L5 aren't simple AI technology applications but must combine with professional skills, network and data application scenarios, meeting stringent telecommunications network reliability requirements.

We've taken crucial steps with leading global operators toward autonomous networks, especially the L3-L4 transition. Notably, in Denmark with TDC, we achieved the world's first TMF-certified autonomous network 4.0 application scenario for network intelligence. Single-site power savings amount to thousands of kilowatt-hours. Considering varying electricity prices at different times, the monetary savings from just one or two sites are significant - imagine the benefits when deployed network-wide.

TMF is the autonomous network standard setter and most authoritative third-party certifier. Ericsson's Danish case is globally the first and currently only one achieving TMF L4.0 certification. Similar stories are unfolding globally. In Saudi Arabia, we work with operators on intent-driven data optimization, precisely identifying all network performance changes and trends, enabling proactive intervention and optimization.

This achieved 27% improvement in average user experience rates. In Malaysia, with DNB's network serving six operators in a complex shared network scenario, we helped elevate it to 4.0 level, achieving autonomous and intelligent network slice deployment.

In Spain, not all networks have reached 4.0 - few have due to complexity and massive investment requirements. However, we see excellent results at every step. We worked with local operators on user complaint handling, intelligently correlating and analyzing complaint-related network factors while generating proactive solutions, reducing user complaints by 14%. This scenario reached level 3.11. This is just one example of autonomous networks being widely deployed globally with accelerating pace and obvious benefits.

Returning to Ericsson's activities: autonomous network-telecommunications integration presents enormous opportunities and challenges requiring unified standards and open ecosystems - not achievable by any single company.

We're pleased that Ericsson has actively promoted and consistently built very open ecosystem systems. We've become the largest owner in the wireless autonomous network industry ecosystem, with over 63 published autonomous network applications and over 61 influential business partners. The open source community for autonomous network application developers has nearly 2,000 members, with 32 testing environments providing not just open source community support but testing facility support for smaller or independent application developers globally.

Finally, we're delighted to participate in the International Fair for Trade in Services. Ericsson has been in the Chinese market for over 130 years, being the only company participating in China's complete telecommunications network construction from 1G to 5G, providing extensive service support and innovation momentum.

Looking forward, Ericsson firmly believes only open market environments and unified standards can achieve healthy competition and prosperous development in the telecommunications industry. Only by achieving this can we succeed in new productive forces and China's and global digital intelligent transformation. Regarding this, Ericsson will continue combining global experience with Chinese practice in autonomous network development, making our contribution.

Thank you very much for your attention!