Why autonomous networks could be a game-changer for emerging markets

Details: Manny Pham; Category: Operators; 13 August 2025; 14831 views

China Mobile (CM) was recently hailed for becoming a high-level autonomous network in only six years. The operator’s efforts were not wasteful and have showcased results that all operators, not just emerging market players, should strive for.

According to analyst house STL Partners, CM had achieved Level 4 autonomy resulting in a 50% increase to user download speed in congested areas and an 80% cut to network faults.

“These are only a few metrics that operators aim for to cut costs and increase efficiency,” says David Martin, Associate Senior Analyst at STL Partners. “More efficient networks mean you release capacity, improve customer experience, cut out costs and free up resources for more innovation.”

The desire for such boosts is more prevalent in developed markets where margins and revenues are shrinking and flattening. However, in emerging markets, operators still have a lot of ground to gain.

Martin says that does not mean developing market operators should take their eyes off developments in autonomous networks. Adopting the technology early can tackle constraints that developed counterparts are already facing in their journeys to automation, such as skills shortages and capex constraints - all against surging demand for connectivity in their nascent markets.

Understanding the levels

A fully autonomous telecoms network has yet to come to fruition. To use the TM Forum’s Autonomous Network initiative as a measure, operators can be ranked on a scale from Level 0 to Level 5.

Level 0 is a network with manually performed network operations with minimal to no automation. Meanwhile, Level 5 is a fully autonomous network that can manage and optimise itself without any human involvement.

“There’s a temptation to focus on the big picture of Level 5,” Martin says. “But you should see it as something that helps you address real pain points in specific network domains now.”

Lessons from China Mobile

China Mobile has made rapid progress to achieve Level 4 from Level 1 in mere six years. A Level 4 network shows self-optimising capabilities and can handle most operations autonomously, but there is human intervention if needed.

Martin says CM began with “low-hanging fruit” - clear pain points in certain regions or domains - and solved them not just in isolation but by addressing the underlying processes. This included using AI and machine learning to predict when a fault was likely to occur and automating the fix before it happened.

“You start small, you get measurable results like saving money or improving Net Promoter Scores, and then you build from there,” Martin says. “That’s how you get the momentum to go from one domain to cross-domain, and eventually to holistic autonomy.”

Opportunities and challenges for emerging markets

Kitesh Bhayani, Research Director at IDC, observes that awareness of the need for network automation is growing in emerging markets, with the Middle East and Africa in particular reporting relatively high adoption compared to some global peers.

He points to several common early applications, including network monitoring and anomaly detection to identify and address faults before customers are affected, predictive maintenance using historical and live data to anticipate issues such as weather-related damage, intelligent traffic routing to manage congestion during peak periods or major events, and automation in customer support and billing to cut costs and improve responsiveness in multiple languages.

However, he also highlights significant obstacles, from shortages of AI and automation expertise to the complexity of managing multi-vendor environments, entrenched legacy systems, and organisational inertia. Much of the older equipment in these markets is incompatible with modern APIs and often deployed in challenging locations, making upgrades expensive and the return on investment uncertain.

To address the skills gap, some operators are partnering with universities to develop tailored courses, while others outsource to major integrators such as TCS or Accenture. Yet Bhayani stresses that even with vendor support, automation is far from a plug-and-play process and requires adaptation to local contexts.

The vendor factor

For Wayne Lotter, Head of International Networks at Telstra International, simplifying the vendor landscape is essential for automation. “If you have hundreds of variants of technologies, whenever you do an action you have to perform it in a model first and in a hundred others - that’s hard for automation,” he says.

Telstra works closely with strategic partners like Nokia and Ciena to keep its technology stack streamlined. “In emerging markets this is crucial because you need to reduce complexity and make it easier to deploy automation reliably,” Lotter explains.

Martin adds that CM’s success was partly due to close collaboration with its major vendors, Huawei and ZTE, on integrating automation even in proprietary systems. “That kind of cooperation can be harder in other developed markets, but it’s a template for how to make progress,” he says.

Resilience in unpredictable environments

Lotter says that in Southeast Asia, automation and AI are vital for resilience. “You not only have disruptive weather but rapid industrialisation, which can lead to building works accidentally cutting off telecom infrastructure,” he says. “By tapping into advanced network technologies, operators can make their networks more resilient. You can pre-empt and detect outages which can be really crucial in those markets. During an outage, traffic can be re-routed before literal disasters strike - ensuring it’s business as usual and customers are none the wiser.”

He likens it to dark automotive factories in China - autonomous factories that are literally kept in the dark as robots on assembly lines construct cars - enabling vendors to cut costs and pass savings on to consumers.

The energy efficiency opportunity

Bhayani and Lotter agree that AI-driven energy management is another strong early target. In the RAN, AI can power down equipment during low usage and reactivate it when needed. Bhayani says even a 5–10% improvement in efficiency “is significant when you consider how much of an operator’s cost base the RAN consumes.”

Lotter notes that Telstra has deployed AI in parts of its network to fine-tune energy usage with “great results” and expects the benefits to grow as demand for data and cloud services increases.

Start small, scale fast

Martin’s advice for emerging market operators is clear: “Don’t think of it as investing in autonomous networks - think of it as investing in making your networks more automated and efficient. It’s not a luxury. It’s a necessity to reduce your cost base and increase your margins.”

Bhayani reinforces the “crawl, walk, run” approach. “Start with low-risk use cases and gradually increase the complexity. That allows for learning, adaptation, and upskilling of your workforce while showing tangible benefits to the business,” he says.

Lotter adds that the aim is not to replace people but to augment their work. “Autonomous networks aren’t about replacing people - they’re about enabling networks to respond faster, smarter, and more consistently than humans can alone. In markets where demand is surging and resources are limited, that can be the difference between leading and lagging.”

As CM’s experience shows, progress can be rapid. Six years ago it was at Level 1. Today it is close to Level 4 across much of its network. For emerging market operators willing to start small, prove value, and build capability, the leap to higher autonomy may come sooner than expected.