How telcos are reshaping network economics in the AI era

The telecom industry is undergoing a structural shift as traffic patterns, AI workloads and ultra-low latency services redefine how networks are built and funded. Speaking on these changes, Randeep Singh Sekhon, CTO, Bharti Airtel, in conversation with Jaijit Bhattacharya, Founder and President, Centre for Digital Economy Policy Research, explained that the nature of telecom traffic itself is evolving.

Sekhon noted that AI has not fundamentally increased traffic yet, but it has changed where traffic flows. “Search has not disappeared,” he said. “It has simply moved from traditional search engines to AI platforms like ChatGPT, Gemini and Perplexity.” For now, AI-driven queries still behave like conventional data traffic, but this will change as AI becomes more embedded into real-time applications.

AI at the edge will change infrastructure economics

A key inflection point, according to Sekhon, will come when AI moves from the cloud to the network edge. Technologies such as smart glasses, facial recognition and real-time decision systems will require AI models to operate closer to users.

“When AI sits at the edge, latency becomes critical,” Sekhon explained. “You are no longer just downloading data. You are making decisions in real time.” This shift will increase the need for edge computing, low latency networks and intelligent routing, pushing telecom operators to rethink both capital expenditure and network design.

India’s data consumption continues to grow rapidly, driven by OTT platforms, social media and high-resolution video. Content that once streamed at 720p is now routinely consumed in 4K, increasing overall data loads.

While downlink traffic still dominates, uplink performance has become strategically important, especially for applications involving cloud access, AI interaction and live services. Sekhon highlighted that customer experience now guides investment decisions more than raw capacity expansion. “Our two pillars are very clear,” he said. “Customer experience and cost to serve.”

Energy efficiency and AI-led cost optimisation

Energy is now the single largest operational cost for telecom networks. Airtel is increasingly using AI and machine learning to optimise energy consumption without compromising coverage.

One practical example is AI-based capacity management. During low-usage hours, Airtel selectively shuts down excess network capacity while keeping coverage intact. “That is our biggest real-world monetisation use case,” Sekhon said.

The operator is also expanding the use of solar-powered base stations and battery storage, allowing energy generated during the day to be used during peak hours. AI models, Sekhon added, are now far more efficient than earlier automation systems in managing these trade-offs.

AI is also playing a critical role in telecom security. Airtel has deployed real-time systems to sandbox suspicious URLs, detect fraudulent links and block scam attempts before they reach customers.

Sekhon shared a real-life example involving call forwarding fraud triggered by a malicious link. “Even very educated users can fall for it,” he said, underlining why real-time AI-based detection and multi-factor authentication are essential.

Financial fraud remains the highest-risk area, and Sekhon stressed the need for closer collaboration between telecom operators, banks and OTT platforms to secure SMS, voice OTPs and messaging channels.

On satellite telephony, Sekhon took a pragmatic view. While satellite-based Wi-Fi and SOS services will arrive in India, he believes terrestrial networks will continue to dominate. “India is largely well served by fibre and mobile networks,” he said. Satellite services will mainly fill niche gaps such as remote locations and emergency connectivity, unlike countries with large unconnected landmasses.

The future of telecom economics

As networks become more intelligent, investments are shifting towards AI-led automation, edge computing, transport modernisation and cloudified cores, rather than linear capacity expansion. Sekhon emphasised that automation investments are evaluated through efficiency gains, whether it is faster field operations, reduced energy usage or improved fault resolution.

In the AI era, telecom economics is no longer about adding more hardware for more users. It is about using intelligence to deliver better experiences at lower cost, while keeping networks secure, sustainable and resilient.

Written by- Preeti Anand