Can AI help telcos achieve net zero?

The price of energy may have fallen since the peaks of 2022, but improving network energy efficiency is set to remain a live topic for telcos attending MWC 2024. And for good reason. Energy costs still constitute between 20% and 40% of a telco’s operational expenses, according to the GSMA.

But although cost reduction is a major incentive, it isn’t the only one. Many of the world’s largest communications service providers (CSPs) have set targets to achieve net zero emissions by 2040, with machine learning (ML) and AI among the tools they are sharpening to reduce greenhouse gas (GHG) emissions.

When it comes to energy efficiency, the telecoms industry has already spent much of the past decade putting machine learning, artificial intelligence (AI) and data analytics to work on understanding and improving network energy usage, as we explored in our report The sustainable telco: engineering networks for net zero.

Verizon, for example, has developed a data analytics platform which provides a virtual representation, or digital twin, of its physical network to analyze and predict the performance, cost and efficiency of its network sites and equipment.

“We were able to gather the power consumption data for all of the cell sites and see some crazy anomalies,” said Michael Raj, Vice President of Network Enablement (AI & Data), at Verizon, in an interview for the report. As a result, “we could question why this costs $1,000 versus the same configuration elsewhere, which only costs $500…and identify specific problems related to that outlier, as well as make recommendations for how to remediate those situations when found in the future.”

Verizon was also able to visualize and compare how different vendors’ equipment consumes energy in the field as opposed to in a lab, and uses the results to push for better performance. “Let’s say for 100 gigabytes of capacity vendor A proves to be 35% more efficient in practice than vendor B, now we can push vendor B to create a more energy efficient solution,” according to Raj.

Other initiatives include using ML to activate sleep mode on a RAN network. Indian operator Jio, for example, developed an AI-driven Cognitive Platform to draw on RAN data from multiple vendors to intelligently shut down the RAN power amplifier for microsecond intervals by detecting when there is no traffic.

Telenor meanwhile developed a global program for energy efficiency, using analytics tools from a variety of vendors to gain more visibility into site and network operations. In some cases, the company found a gap of 30% to 35% between its spending on energy and the amount of energy it needed to operate its active network. The company therefore focused on bringing energy consumption in line with what it required to generate signals, transfer data and serve customers. The strategy soon reaped benefits: Within less than a year of starting the program it saw a 2% reduction in energy consumption, while at the same time increasing network capacity to support a 35% rise in traffic.

The Scope 3 challenge

Reducing energy consumption brings clear environmental and financial benefits to telcos. However, it is not enough to help them achieve net zero greenhouse gas emissions.

Using energy to run networks falls under Scope 2, or direct emissions. Yet by far the biggest contributor to GHG emissions for communications service providers is Scope 3, or indirect emissions, including those from the production, use and disposal of purchased goods and services.

They are also on the rise, according to 53% of 68 sustainability executives, working for 50 principally large telecoms operators worldwide, who took part in a TM Forum survey for our recent report The sustainable telco: navigating the maze of scope 3 emissions. Indeed, Scope 3 accounts for more than 90% of GHG emissions, according to more than one third of those surveyed for the report, and more than 75% by almost a third more.

Telcos are keen to use technology, including AI, to address Scope 3 emissions. Multiple challenges, however, exist, particularly when it comes to accessing the data they need to understand the scope of the problem, never mind feeding ML and AI systems.

A primary obstacle is that Scope 3 emissions are generated beyond a telco’s direct control through activity up and down the supply chain. This makes it difficult to capture the multiple data points they need to build an accurate picture of their Scope 3 emissions. In addition, many telcos describe a lack of industry-specific reporting standards for use by suppliers, resulting in a paucity of accurate, standardized data that AI systems feed on.

AI-driven customer choice

Telcos are working with suppliers and each other to tackle the issue of how to access standardized data, as well as other factors that impact Scope 3 emissions, such as the recycling and reuse of equipment. They are also weighing up how the future shape of networks and initiatives such as infrastructure-sharing and Open RAN could impact their Scope 3 carbon footprint reduction.

In the meantime, CSPs also have to meet the demands of regulators, investors, employees and customers to reduce GHG emissions. And despite broad-reaching issues with access to accurate, standardized data, AI is proving useful when it comes to solving individual Scope 3 challenges.

Vodafone, for example, championed a TM Forum Catalyst alongside Carbon Footprint, which won this a Moonshot award at DTW23-Ignite.

The project sets out to tackle the problem of measuring Scope 3 by helping CSPs determine value chain emissions data and easily transpose that data to a product catalog. The aim of the Digital Carbon Footprint Optimization catalyst is to use predictive insights based on AI / machine learning to help customers visualize the CO2 footprint of different products and services.

They can then decide if they want to opt for more environmentally friendly usage patterns, says Jonas Blume, who until recently was Business Manager to the CIO, Vodafone Germany. “This transparency is really important to consumers and it’s getting more and more important every year,” explains Blume.

“With the Catalyst we wanted to show how to make customers aware of the more environmentally friendly solutions they can choose from their CSP.”

Vodafone also wants the tool to enable its own product managers to take into account the CO2 footprint of a service in a catalog and then “use intelligent search tools or generative AI to build alternative lower carbon product offers”.

The service considers the carbon footprint of a device as well as how usage – of a mobile application, for example – impacts the battery usage of a handset. Network performance is the third metric. This is important because “if you don’t have a good connection, your phone consumes more energy as it tries to reconnect,” says Blume. A data and analytics platform then creates personalized insights for consumers, using a new TM Forum Open API to display the data.