AT&T and Deutsche Telekom Pioneer Agentic AI and Sensing in 6G Evolution

Image: Illustrative image · Roger Wollstadt · CC BY-SA 2.0 · Source

AT&T and Deutsche Telekom are advancing 6G development by integrating agentic AI and Integrated Sensing and Communication (ISAC) with current 5G networks, emphasizing a continuous evolution of connectivity and computing.

At the TIP FYUZ 2026 event in Seattle, leading telecom operators AT&T and Deutsche Telekom outlined their strategies to accelerate 6G technology by leveraging existing 5G infrastructure. They highlighted agentic artificial intelligence (AI) and Integrated Sensing and Communication (ISAC) as central to this evolutionary path.

Rob Soni, AT&T’s Vice President of RAN Technology and TIP board chair, described 6G not as a disruptive leap but as a progressive enhancement of connectivity. He identified ISAC as a compelling focus for 6G development, with AT&T already implementing it atop current 5G networks. Earlier demonstrations at the AT&T Stadium showcased sensing capabilities, indicating practical steps toward this integration.

However, Soni acknowledged the challenge of business models supporting ISAC, particularly outside defense and military applications, noting the importance of clarifying who will bear associated costs. Due to the substantial data involved, AT&T is experimenting with server placements adjacent to baseband units to manage ISAC efficiently, anticipating future consolidation into a single server architecture as commercial processors evolve.

AT&T anticipates building around 1,000 data centers in the US over the next few years, predominantly near existing fiber infrastructure, to support the advanced demands of 6G including sensing, satellite, Wi-Fi, and mobile compute capabilities. Improving cell-edge performance remains critical, given that current speeds at the periphery of cells fall short of requirements for autonomous vehicles, a key use case.

Deutsche Telekom’s Senior Vice President of Group Technology, Thomas Lips, approached the developments with cautious pragmatism. While acknowledging the focus on sensing and compute, he emphasized that customers still demand connectivity improvements such as uplink speeds, capacity, and reliability at the cell edge. He called for operators to address their customers' immediate needs and to develop 6G's AI infrastructure that supports multiple cooperative autonomous agents rather than ad hoc AI additions.

Deutsche Telekom has already implemented a multi-agent system that autonomously optimizes network performance during major events, while AT&T is developing tailored telecommunications AI models, including a substantial 18 billion parameter model for radio access networks (RAN). Furthermore, AT&T is advancing root cause analysis of call drops through AI agents using open-source frameworks, aiming toward proactive detection without customer input by 2027.

Both operators see 5G-Advanced as a valuable platform for ongoing testing and gradual integration of sensing, computing, and connectivity, laying the groundwork for fully unified 6G services. They encouraged collaboration among smaller vendors and operators to ensure inclusive progress across the industry. This approach highlights that 6G is poised to be a comprehensive ecosystem upgrade, spanning hardware, software, and business models rather than a simple interface update.