SK Telecom’s ATHENA Vision for AI-Native 6G Networks

SK Telecom has been publishing annual white papers outlining its perspective on the evolution towards 6G. Earlier documents discussed the lessons learned from 5G deployment and introduced the concept of AI-driven telecom infrastructure. In its latest white paper, the company moves further along that path by presenting a vision for future network architecture called ATHENA.

Rather than focusing solely on individual technologies, the document examines how the overall telecommunications infrastructure may evolve in the coming decade. It looks at the architectural foundations required to support emerging services, the growing role of artificial intelligence, and the operational challenges that operators will face as networks become more complex and service requirements expand.

The white paper argues that the telecom industry is entering a phase where networks must simultaneously support two directions of innovation. The first is the increasing use of AI to optimise and automate network operations. The second is the need for networks themselves to provide the computing and connectivity capabilities required for large scale AI services.

To address these trends, SK Telecom proposes a long term infrastructure evolution built around six guiding principles: AI-native networking, cloud-native design, ubiquitous connectivity, openness, zero trust security, and customer centric operation. These ideas form the conceptual foundation of the ATHENA architecture.

The name ATHENA itself is derived from six key attributes of the proposed network architecture: AI, Trust, Hyper-connectivity, Experience, opeN, and Agility. The architecture is designed to integrate these characteristics across the main components of the mobile network, including the radio access network, the core network, the transport layer, and a new data platform designed to extract value from network data.

One of the central themes in the white paper is the idea of AI-native networking. In this model, AI is not simply an external optimisation tool but becomes embedded in the architecture of the network itself. Networks are expected to evolve into intelligent systems that continuously analyse operational data, predict potential problems, and automatically adjust configuration and resource allocation.

At the same time, the network must also support AI workloads. This means telecom infrastructure will increasingly host computing resources capable of running AI training and inference workloads. By integrating computing capabilities into the network infrastructure, operators could potentially support new services while making better use of existing network resources.

The radio access network is a key focus area in the proposed architecture. SK Telecom describes an evolution towards AI-native RAN, where open interfaces, virtualisation, and edge computing resources enable a more flexible and intelligent radio network. The architecture introduces the idea of base stations evolving into intelligent nodes capable of running both communication functions and AI workloads.

In this model, general purpose hardware equipped with specialised processing units can simultaneously support radio access processing and edge AI applications. This allows the same infrastructure to deliver connectivity services while also supporting applications such as autonomous systems, immersive media services, and other latency sensitive workloads.

Automation also plays an important role in the RAN evolution. The architecture relies on orchestration systems and intelligent controllers capable of analysing real time network data and automatically optimising parameters such as load balancing, coverage, and energy efficiency. Over time, the goal is to reduce the level of manual intervention required in network operations.

The core network is also expected to evolve significantly. SK Telecom envisions a future core architecture referred to as xCore, which extends the cloud native principles introduced in 5G. Network functions are expected to run as microservices that can be dynamically deployed across distributed cloud environments.

AI is again central to this evolution. The white paper discusses the use of AI agents capable of analysing operational data, predicting network behaviour, and managing resources autonomously. These agents could support tasks such as fault prediction, traffic optimisation, and automated recovery from failures.

Another important element in the architecture is the concept of network exposure. By exposing network capabilities such as quality of service, location information, and security features through standardised APIs, operators can enable new service ecosystems. This approach is aligned with the broader industry trend towards network platforms that allow developers and enterprises to build services that interact directly with network capabilities.

Transport networks are also expected to undergo significant transformation. SK Telecom proposes an AI-native converged transport network that integrates optical and packet transport layers while enabling intelligent resource management. The aim is to simplify the network architecture while improving scalability, automation, and efficiency.

Security is addressed through a zero trust approach that assumes no component of the network can be implicitly trusted. Instead, all connections and interactions are continuously verified. This approach becomes increasingly important as networks become more open, more virtualised, and more interconnected with external services.

A new architectural component introduced in the white paper is the Data Insight and Value Engine, or DIVE. This platform is designed to collect and analyse large volumes of network data in order to support operational optimisation and new services. The platform is expected to operate across both edge and cloud environments, providing analytics capabilities that can support AI driven operations as well as data based services.

Taken together, these architectural components reflect a broader shift in how telecommunications networks are expected to evolve. Instead of focusing solely on higher speeds or new radio technologies, the emphasis is increasingly on intelligence, automation, and the integration of computing capabilities within the network.

While the exact shape of 6G networks will ultimately depend on standardisation outcomes and industry collaboration, the ideas presented in SK Telecom’s ATHENA architecture provide an interesting view of how operators are thinking about long term infrastructure evolution. The white paper highlights how future networks may become deeply integrated platforms that combine connectivity, computing, data analytics, and automation to support a new generation of digital services.

For readers following the progression of SK Telecom’s 6G white papers, this document marks a shift from early technology exploration towards a more concrete architectural vision. It reflects the growing recognition that the success of future networks will depend not only on new radio capabilities but also on how intelligently the entire telecommunications infrastructure is designed and operated.

Related Posts: 

Posted by on

Comments

Post a Comment