India's mobile operators will need substantially larger network capacity — bigger telecom "pipes" — to carry the traffic generated by the AI boom, according to the director general of the Cellular Operators Association of India (COAI), the industry body representing the country's telcos.

The warning lands as AI applications shift from text-based chatbots to bandwidth-heavy workloads, pushing data consumption patterns beyond what current network dimensioning anticipated. At the same time, the COAI chief flagged spectrum availability as a persistent obstacle to expanding that capacity, signalling that the industry sees airwaves — not just backhaul or compute — as the binding constraint on India's AI readiness.

Why does AI change network planning?

AI services differ from traditional mobile broadband in how they consume connectivity. Interactive AI applications rely on low-latency, always-on sessions rather than bursty downloads, and as generative tools move into voice, video and multimodal formats on consumer devices, uplink and downlink demands grow in tandem.

That shift puts pressure on operators to upgrade radio capacity, transport networks and core infrastructure. The COAI chief's point is direct: the AI economy runs on connectivity, and if the pipes stay at today's dimensions, the growth in AI usage will hit a network ceiling.

For India — one of the world's largest mobile data markets by volume — the stakes are particularly high. The country's operators have spent recent years rolling out 5G while still carrying the bulk of traffic on 4G, and any step-change in per-user consumption from AI services would land on networks already balancing capital expenditure against pricing pressure.

Where does spectrum fit in?

Spectrum remains the hurdle, according to the COAI chief. Operators have repeatedly argued that the cost and availability of airwaves in India constrain how much capacity they can economically add. Expensive spectrum translates into thinner holdings per operator, which in turn caps how much traffic a network can carry at given quality levels.

The industry body's position implies that if AI-driven traffic growth materialises as expected, the current spectrum assignment framework will need revisiting — through additional auctions, revised pricing, or faster release of bands — before operators can scale capacity to match demand.

What comes next?

The COAI chief's comments frame capacity building as a precondition for India's AI ambitions rather than a downstream consequence of them. The practical question for policymakers and operators alike is whether spectrum supply, network investment cycles and AI adoption curves can be aligned in time to keep the pipes ahead of the traffic.