Artificial intelligence (AI) and data centres are booming industries, ones that could potentially lead to everything from medical breakthroughs to solutions to the triple planetary crisis of climate change, nature and biodiversity loss, and pollution and waste. To give just one example, the United Nations Environment Programme’s (UNEP’s) International Methane Emissions Observatory is using AI to analyze satellite data and notify companies and countries about emissions of methane – a powerful climate-warming gas. However, the expansion AI and its associated infrastructure has growing impacts on the environment, people and economies. We recently spoke to Hongpeng Lei, Chief of the Mitigation Branch in UNEP’s Climate Change Division, about how the world can harness the benefits of AI without damaging the environment.
What are the biggest environmental concerns around the explosion of AI and data centres?
Hongpeng Lei (HL): It is well established that data centres place growing demand on electricity systems and can also affect water resources – depending on their design, cooling technology and location. While energy demand is rising across all AI data centres, water impacts vary greatly, with some relying on water-based cooling, particularly in hot or water stressed regions.
The spike in power use is particularly time‑critical, given that the world is rapidly approaching global temperatures that risk exceeding the aspirational 1.5°C limit set by the Paris Agreement.
To give some numbers, in 2024, global data centres were estimated to have consumed 415 terawatt-hours of electricity – approximately 1.5 per cent of global electricity. This consumption is predicted to double by 2030. And, of course, when this power is generated from fossil fuels, we have a major emissions issue.
Fundamentally, sustainability is not a constraint: it is a pathway to resilient, competitive digital infrastructure.
What can we do to foster the sustainable expansion of data centres?
HL: A key measure as we shift towards full decarbonization of the global economy is efficiency – of both energy and resources. Efficiency is important not just to reduce emissions, but to reduce costs for communities living close to power centres and ease strain on power grids. According to Bloomberg, wholesale electricity prices rose up to 267 per cent in areas near large data centres over the last five years. This has an obvious financial impact on local economies. And we could be looking at increased outages: the International Energy Agency (IEA) predicts that planned AI data centres could account for 5-10 per cent of peak electricity demand in major European markets.
Efficiency measures can make a massive difference, often faster and at a lower cost than expanding power supplies.
What kind of measures are we talking about?
HL: There are many tools at our disposal. One of the most effective levers is sustainable procurement, which influences how data centres and their equipment are designed and operated, helping to ensure that the equipment and installation are sustainable, energy-efficient and climate friendly.
Procurement specifications can set clear expectations on energy efficiency, cooling performance and operational best practices, reducing electricity demand and associated costs from the outset. They can also promote transparency through common performance metrics.
These are all areas in which the UNEP-led United for Efficiency initiative is providing support to governments, including by developing sustainable procurement guidelines for data centres and computer servers. These guidelines support governments, companies and data-centre operators in applying these measures consistently and at scale.
Can AI and digitalization themselves provide solutions to these problems?
HL: Absolutely. We are seeing exactly this with the Demand-Driven Electricity Networks Initiative (3DEN), which is run by UNEP and the IEA, supported by the Italian government. The initiative backs smart technologies that allow greater efficiency in power systems and make it easier to integrate renewable sources into grids, which can deliver power to millions more people without increasing emissions.
For example, 3DEN backed a consortium in India to develop a digital twin for electric distribution grids, which enhanced operational reliability, improved outage management and reduced costs for 20,000 consumers.
According to the IEA, digitalization that allows precise management of energy systems across sectors could cut global energy consumption in buildings alone by up to 10 per cent by 2040. So, these kinds of efforts are well worth pursuing.
What would you say to governments who are still hesitant to implement measures to increase the sustainability of data centres?
HL: AI data centres are here to stay. They have unleashed and can unleash all sorts of benefits. But if their expansion is left unmanaged, it can drive up local electricity demand, push up power prices for households and businesses, increase pressure on power and water systems, and raise emissions. Governments, local and national, should remember that managing power and water use more efficiently is essential to protecting electricity affordability, maintaining system reliability and attracting investment. Fundamentally, sustainability is not a constraint: it is a pathway to resilient, competitive digital infrastructure.
On 26 January, the world will mark the International Day of Clean Energy. The observance is designed to raise awareness and mobilize action for a just and inclusive transition to clean energy for the benefit of people and the planet.
