How Ericsson Is Teaching Networks to Run Themselves

Why It Matters

Achieving Level‑4 autonomy will reduce operational costs and unlock new AI‑driven services, while Ericsson’s role as an orchestrator between operators and hyperscalers determines its competitive edge in the evolving 5G ecosystem.

Key Takeaways

• Operators aim for Level‑4 network autonomy by 2030.
• Human‑in‑the‑loop remains essential for near‑term autonomous networks future.
• AI integration escalates as networks progress from Level‑2 to Level‑4.
• Ericsson leverages 5G macro‑network slicing for enterprise and private deployments.
• Collaboration with hyperscalers crucial to build AI‑centric mobile infrastructure.

Summary

The interview spotlights Ericsson’s roadmap for autonomous networking, outlining how the company envisions mobile operators moving from manual management toward intent‑based, AI‑driven networks by the end of the decade.

Around 80 % of carriers rate themselves early on the change curve but set a target of Level‑4 autonomy by 2030, which still requires a human‑in‑the‑loop for decision oversight. Progressing from Level‑2 to Level‑4 will demand increasingly sophisticated AI, while 5G standalone and network slicing enable macro‑network resources to serve both private‑edge and enterprise deployments.

Ericsson cites mission‑critical use cases—full‑coverage, low‑latency services—as proof points, and highlights its recent turnaround in cloud software and services that has boosted market share and customer trust.

The push toward AI‑centric connectivity positions Ericsson as a bridge between telcos and hyperscalers, making its technology stack essential for future AI applications across industry, consumers, and society.

Original Description

How close are we to truly autonomous networks - and what’s actually standing in the way?

In this conversation, Ericsson’s Jenny Lindqvist breaks down where operators really are on the autonomy curve, why most are still early in the journey, and what it will take to move toward Level 4 autonomy by 2030. The shift isn’t just technical - it’s about complexity, trust, and the growing role of AI in managing increasingly dynamic networks.

The discussion also explores a bigger inflection point: networks are no longer just built for connectivity - they’re being rebuilt for AI. As hyperscalers push into the space and enterprise use cases expand, the question becomes less about competition and more about orchestration. Who controls the stack, and how do these systems actually work together?