AI applications will force telecom networks to rethink uplink, spectrum and power: Ericsson India MD

AI applications could push telecom operators to redesign their networks around uploads rather than downloads, as more data flows from users to the cloud and AI infrastructure, Ericsson India managing director Nitin Bansal said.

As the technology moves beyond chatbots and into devices such as AI-enabled glasses and other applications that continuously send data to the cloud, the uplink side of mobile networks could become increasingly important, Bansal told ET AI at the India Mobile Congress (IMC) 2026.

“For example, if you're using AI glasses, you need an enhanced uplink because you're sending a lot of data, photos and videos on the uplink,” Bansal said.


Ericsson refers to this shift as part of “networks for AI”, building network capabilities around the requirements of AI applications. The company is also looking at the other side of the spectrum, or “AI for networks”, where the technology is used to improve network performance and efficiency.

Also Read: India has 12-18 months to get its 6G contributions into global standards: Nokia

For telecom operators, the latter could translate into better use of spectrum, lower energy consumption and the ability to support more users without a proportional increase in network resources.

“One is spectral efficiency,” Bansal said, adding that AI can help operators use spectrum more efficiently and “carry more users” with the same amount of spectrum.

Reducing power consumption

The other major opportunity, as per Bansal, is reducing the power consumed by network equipment, which could also lower operating costs for telecom operators. “It’s about performance, capacity, and efficiency.”

The distinction between AI for networks and networks for AI could also become important as operators look for ways to monetise investments in AI-ready infrastructure.

Bansal said applications requiring enhanced network capabilities could potentially lead to better experiences and greater consumption of connectivity.

“People will consume more when they have a good experience,” he said, pointing to enhanced uplink capabilities for AI-enabled devices as one example.

The comments come a day after Ericsson said it had produced more than 15,000 antennas and 25 million components in India since July 2025. The antennas manufactured in India were exported to Europe and other global markets.

In the announcement, the company said its antenna portfolio is being developed to support the transition towards more intelligent and programmable networks, including improving coverage and capacity, strengthening uplink performance, using spectrum more efficiently and reducing network energy consumption.

Also Read: India could need 3 GHz spectrum for 6G as AI pushes up network demand: GSMA

Ericsson’s India claims its manufacturing push is tied to the changing requirements of networks as operators move towards 5G Standalone, a 5G architecture that uses a dedicated 5G core network rather than relying on existing 4G infrastructure, AI-driven operations and eventually 6G.

Bansal said India’s large-scale 5G rollout puts the country in a relatively strong position and pointed to the growing need to connect the data centres to support AI workloads.

“Data centres need to be connected,” he said, adding that 5G could provide a “secure, trusted and robust layer” for connecting data centres and enabling the flow of data.

The evolution towards 6G is expected to bring these trends together. While 6G standards are still being developed, Bansal identified 5G-Advanced, network slicing and AI as three key building blocks.

India is also making a “tremendous contribution” to 6G standardisation, he said, adding that the country is “on the right track” in its ambition to play a larger role in next-generation networks.