e& UAE and Khalifa University have unveiled a new white paper outlining a blueprint for AI-native 6G networks, positioning intelligence as a core component of future connectivity rather than an add-on feature.
Titled “6G AI-Native Networks: Architecture, Intelligence, and the Path to Autonomous Connectivity,” the paper introduces a fundamentally new network architecture that embeds learning, reasoning, and autonomous decision-making directly into network operations. The framework is designed to contribute to global standardisation efforts, including IMT-2030 and 3GPP Release 21+.
A central concept in the proposal is the introduction of a dedicated “AI-plane,” added alongside traditional user, control, and management planes. This AI-plane enables continuous sensing, learning, and autonomous action across the network, spanning radio, core, and edge environments.
The architecture also incorporates distributed AI agents, closed-loop autonomy, digital twin integration, and AI lifecycle orchestration. These capabilities allow networks to self-optimise, predict demand, and manage performance in real time across multi-vendor and multi-domain environments.
The white paper outlines five core pillars of AI-native 6G: pervasive AI and machine learning, distributed cloud-edge computing, advanced sensing technologies, open programmable architectures, and sustainability-focused design. Together, these elements position 6G as a cognitive infrastructure capable of powering advanced use cases such as immersive experiences, smart cities, and intelligent transportation systems.
To support standardisation, the framework introduces measurable AI-specific KPIs, including decision latency, learning accuracy, and energy efficiency per AI inference—providing a structured approach to evaluating autonomous network performance.
The collaboration highlights the UAE’s ambition to lead in next-generation connectivity, combining academic research with industry deployment to shape global 6G standards and accelerate the transition toward intelligent, autonomous networks.
