- Michal Sewera (Deutsche Telekom) compares the cloud-native 5G network to the “butterfly effect”—a small change in one of hundreds of modules (like a service mesh) can cause major errors down the line, only detectable through KPI fluctuations.
- To handle this complexity, DT is implementing a new AI-based operating model:
- Development phase: AI supports testing, regression, and validation.
- Deployment phase: Automated network configuration via a “network configuration copilot.”
- Operations phase: AI handles service monitoring, root cause detection, and resolution.
- Sewera emphasized the need for a new generation of network engineers who can code and develop software, instead of the traditional silos of 10–20 years ago.
- At Telenor, Terje Jensen stated that their AI + 5G revenue strategy focuses on individual enterprise customers: “custom solutions for specific value.” Typical applications include industrial video analytics, remote driving, and “network sensing”—collecting user movement data from the network infrastructure.
- Telenor has established an “AI Factory” that offers GPU-as-a-service, allowing companies to rent computing power for AI training.
- Andy Corston-Petrie (BT Group) warned that AI investment is expensive and requires a new mindset—combining small, flexible, and distributed models instead of massive AI projects.
- Sewera urged the industry to avoid the “definition fever” that occurred with cloud-native: instead of debating “what is an AI-native telco,” the focus should be on applying AI practically to the network to optimize performance and costs.
📌 Summary: AI is becoming a strategic pillar for telcos like BT, DT, and Telenor in the 5G era, ranging from network automation to new business models. DT is pushing for AI across the entire network lifecycle, Telenor is commercializing GPUs and enterprise AI services, while BT is seeking a distributed, resource-efficient approach. AI not only improves KPIs but also paves the way for new revenue from intelligent infrastructure.
