The managing director of Ericsson India has told The Economic Times that artificial intelligence applications will force telecom operators to redesign three layers of the network stack. He named uplink capacity, spectrum allocation and power consumption as the three pillars facing structural pressure from AI workloads.

The comment, reported in the Indian financial daily, frames AI-driven traffic not as an incremental demand curve but as a planning discontinuity. Operators used to engineer for downlink-heavy consumer traffic. They now face a workload mix dominated by prompts, telemetry and inference returns.

Why does AI change the uplink calculus?

Mobile network planning has historically assumed downlink dominates. Video streaming, web browsing and application downloads sent bytes from the network to the handset. AI applications invert that balance. Models running on the device push prompts upward. Edge nodes relay sensor and inference payloads toward centralized compute. The MD identified uplink as the first engineering line item requiring redesign, per The Economic Times.

Operators must now budget uplink capacity in ways radio planning teams have deferred for a decade. Uplink bottlenecks historically kept small cells from paying off. The shift cascades into backhaul, midhaul and core transport. AI-driven uploads change the economic case for dense radio deployments in Indian cities, which already carry heavy 5G traffic.

Why does spectrum policy return to the agenda?

Spectrum licensing in India and most markets assumes human-mediated traffic patterns. Auctioned bands carry obligations tied to coverage and population, not continuous machine-to-machine inference sessions. The MD told The Economic Times that operators and regulators must revisit band allocation, refarming and shared access for AI session traffic.

The Indian regulator has not formally opened an AI-spectrum consultation. The MD's framing pre-loads the debate with vendor logic. Refarming of mid-band 5G spectrum, dynamic sharing between operators, and licensing of millimeter-wave capacity for fixed AI backhaul all sit inside this conversation, and all touch Department of Telecommunications timelines.

Why is power the binding constraint?

Each AI query burns energy across three points: the device, the radio access network and the data center. The MD identified power as the constraint linking spectrum and uplink decisions to operating cost. Power consumption already sits on board agendas for operators carrying 5G radio loads. AI workloads compound the problem rather than introduce it.

A single inference call may traverse the device radio, a 5G base station, fiber backhaul, and a graphics-accelerated server. Each hop adds kilowatt-hours. Indian operators have spent two years negotiating renewable power purchase agreements to dampen tariff risk. AI traffic raises the marginal cost of every gigabyte transmitted and pushes the energy conversation upstream into radio design.

What does this mean for operators?

Three operational priorities follow from the MD's framing, per The Economic Times:

• Re-engineer uplink budgets and radio planning for prompt and sensor traffic patterns
• Engage with the Department of Telecommunications on spectrum refarming and shared licensing for AI sessions
• Treat power procurement and energy management as network planning inputs, not afterthoughts

Each priority changes a budget line and a regulatory calendar. Each creates an opening for vendor counsel, given the integrated portfolio in RAN, transport and energy systems that sits with the larger equipment vendors.

What changes for vendors?

The MD's framing flows directly from a vendor position. If uplink, spectrum and power re-enter the strategic agenda, suppliers with end-to-end portfolios in radio access, microwave, optical transport and energy management gain leverage. Ericsson sits in that category, alongside its principal RAN competitors. The MD's message carries the weight of a thesis aligned with product strategy, which practitioners should weight accordingly when separating commercial reality from vendor marketing.

What comes next?

The Economic Times did not attach a timeline to follow-up statements or policy engagement. Practitioners should expect Ericsson India leadership to repeat the message at industry forums and to anchor it in upcoming product cycles. Operators hearing the warning should plan a unified response that treats uplink, spectrum and power as one engineering problem rather than three isolated workstreams.