Ericsson and LG Uplus Partnership to embed Voice AI directly into Telecom Networks
Ericsson and LG Uplus have agreed to deepen their long-standing strategic partnership to develop and commercialise network-based voice AI solutions for global markets while accelerating the transition toward AI-native network architectures.
The strategic collaboration represents a fundamental shift in how artificial intelligence is delivered across mobile infrastructures. Rather than relying on standalone, client-side smartphone applications that consume heavy on-device compute or require separate cloud connections, the alliance aims to embed voice AI capabilities directly into the core communications service provider (CSP) network infrastructure.
Decoupling AI Voice Services from Handset Application Boundaries
For years, mobile operators have acted as passive data pipes, routing voice and data packets to third-party OTT (over-the-top) applications. By embedding voice intelligence natively into the network layer via Ericsson’s IP Multimedia Subsystem (IMS), the foundational framework that manages IP-based voice and video sessions, and pairing it with LG Uplus’ commercial ixi-O AI call agent expertise, the partners are creating a network environment that processes conversational logic in real time during the call connection setup.
This architecture enables several high-consequence operational capabilities:
Application-Free AI Invocation: Mobile networks can invoke real-time call screening, automated spam filtering, or live multi-language translation without requiring the caller or receiver to launch or install specific third-party applications.
Reduced Device Dependence: Offloading contextual voice reasoning to secure, network-edge computing nodes ensures lower latency while expanding advanced AI capabilities to low-power hardware, including smart glasses, wearables, and emerging IoT devices.
Enhanced Data Privacy & Trust: Processing voice streams through trusted, high-quality operator channels reduces the need to send raw, unencrypted audio feeds to unvetted external public cloud endpoints.
The Dual Strategy: "AI for Network" and "Network for AI"
Beyond consumer-facing voice services, the strategic framework establishes an operational roadmap for AI-native network evolution, focusing on two interdependent concepts:
Network for AI: Re-architecting 5G Standalone (5G SA) and future 6G environments to support the heavy symmetric bandwidth and ultra-low latency demands of generative AI workloads. Because real-time conversational AI requires fast uplink speeds and minimal multi-server latency (target under 50ms), the underlying radio access network (RAN) and cloud edge must dynamically allocate dedicated network slices to prevent audio jitter.
AI for Network: Applying machine learning algorithms directly to telemetry streams to drive automated network operations. The network monitors its own operational conditions, predicts traffic spikes, dynamically manages energy consumption by powering down inactive sleep-mode radios, and executes self-healing responses to hardware or line outages without manual engineering intervention.
| Operational Strategic Track | Legacy Mobile Operator Model | Ericsson & LG Uplus AI-Native Model |
| Voice Service Delivery | Passive data conduit routing calls to app-based AI agents. | In-Network Intelligence: Native IMS-level AI call processing. |
| Hardware Compatibility | High-end smartphone required for local on-device NPU processing. | Universal Accessibility: Network-level AI works across basic phones, wearables, and IoT. |
| Network Management | Reactive engineering troubleshooting following network failures. | Autonomous Operations: Self-monitoring, predictive fault detection, and energy tuning. |
| Monetization Framework | Static voice and data subscription plans. | Programmable Network APIs: Exposing AI voice capabilities for enterprise co-creation. |
Executive Leadership Perspectives
Per Narvinger, Executive Vice President and Head of Business Area Networks (and incoming President and CEO) of Ericsson, emphasised the structural role of AI in future telecommunications:
“AI-native networks will serve as a key technology foundation for the future of telecommunications. Through our partnership with LG Uplus, we will enhance customer experience and improve operational efficiency based on network-based voice AI solutions and AI-native networks, while jointly exploring new business opportunities and securing new growth engines together.”
Bumshik Hong, CEO of LG Uplus, highlighted the global scalability of their joint commercial framework:
“By combining LG Uplus’ voice AI service experience accumulated through ixi-O with Ericsson’s global network technology capabilities, we will introduce network-based voice AI to the global market, delivering a more intelligent and convenient communications experience. Through this collaboration, we will further shape the global business opportunities for voice AI and ensure that AI-native network technologies contribute to LG Uplus’ mid- to long-term growth and innovation in customer experience.”
The two companies are establishing a structured, long-term cooperation framework to complete initial technology validation and proof-of-concept (PoC) testing within the current calendar year. By combining commercial service deployment insights with global network infrastructure, Ericsson and LG Uplus are laying the groundwork to export programmable, network-based voice AI solutions to telecommunication operators worldwide.
2 Min read