The Network That Thinks: Kraken and the Dawn of Cognitive 6G

Future sixth-generation (6G) networks must move beyond high-speed data delivery to support intelligent, context-aware services. Applications such as autonomous mobility, immersive extended reality (XR), and large-scale sensing require networks that can interpret context, anticipate system dynamics, and coordinate resources according to application objectives. However, current architectures remain fundamentally bit-centric, focusing on reliable bit transport without accounting for the task-level value of information. This article introduces Kraken, a knowledge-centric architectural framework for 6G networks. Kraken integrates three complementary capabilities: semantic communication for transmitting task-relevant information, predictive reasoning for world modeling, and goal-oriented networking for aligning network decisions with application outcomes. These capabilities are realized through a three-plane architecture consisting of Infrastructure, Agent, and Knowledge planes. Through representative scenarios and a system-level evaluation, we demonstrate that knowledge-centric operation can significantly reduce communication overhead while improving latency and task success under dynamic conditions. The article also outlines deployment enablers and key research challenges, positioning Kraken as both an architectural blueprint and a practical pathway toward knowledge-centric 6G systems.