Lead with AI in Autonomous Networks

Why It Matters

AI‑driven autonomy will determine which telcos can scale services profitably while meeting latency, sustainability and cost targets, making it a strategic imperative for the industry.

Key Takeaways

• AI becomes the operating system for cloud‑native telecom networks.
• AI models span edge to core, enabling real‑time link adaptation.
• Ericsson’s AI delivers up to 20% throughput and 75% OPEX gains.
• Deutsche Telekom stresses open, interoperable AI and federated learning.
• Data quality, multi‑vendor integration, and trust hinder AI scale‑up.

Summary

Telecom TV’s panel examined how AI is reshaping autonomous networks as operators move from static hardware to cloud‑native, software‑defined fabrics. With 5G already deployed, the scale and dynamism of services now exceed human‑managed operations.

Both Deutsche Telekom and Ericsson argued that AI is no longer an add‑on but the network’s operating system. AI must handle millions of variables, personalize SLAs in real time, and boost resilience and energy efficiency across the resource, service and business layers.

Ericsson highlighted concrete models—from edge‑embedded link‑adaptation neural nets to centralized R‑apps using graph neural networks and reinforcement learning—showing live gains of up to 20% user throughput, 10% spectral efficiency, and 75% OPEX reduction in cell‑shaping. Deutsche Telekom stressed the need for open, interoperable agents and federated learning to preserve data sovereignty while sharing insights.

The discussion underscored that widespread AI adoption hinges on clean, AI‑native data, multi‑vendor openness, and trustworthy, explainable outcomes. Operators that master these challenges can unlock cost savings, higher performance, and sustainable network growth.

Original Description

As networks evolve from static hardware to cloud-native environments, the complexity has reached a point where manual management is no longer sustainable. Deutsche Telekom’s Ahmed Hafez explains how AI must become the operating system of networks, not just an enhancement, to manage the scale of variables in software-defined fabrics. Ericsson’s Ibrahim Eldeftar details how AI models are being deployed across different network layers, from radio edge optimisation to centralised coordination, delivering measurable improvements, including 20% user throughput gains and 75% operational efficiency improvements in live deployments.

Featuring:

Ahmed Hafez, SVP Technology Strategy and AI in Networks, Deutsche Telekom

Ibrahim Eldeftar, Head of Solution Line Cognitive SW and Services, Ericsson

Recorded June 2026

#AI #AutonomousNetworks #5G #NetworkAutomation #CloudNative #Telecoms #telecomtv