Will AI Help Data Centres Meet Net Zero Goals?
The rise of AI is transforming data centre operations and reshaping the sustainability strategies of the world’s biggest technology companies.
Alphabet, Microsoft, Apple, Meta and Amazon have all set ambitious net zero targets, but the rapid growth of AI workloads is forcing them to confront the energy and emissions challenge head-on.
The latest generation of AI models requires vast computational resources, with thousands of high-performance GPUs running continuously in hyperscale data centres. These facilities demand constant power and advanced cooling, pushing electricity use to unprecedented levels.
According to the International Energy Agency (IEA), global data centre power consumption could exceed 1,000TWh annually by 2026 — similar to Japan’s total electricity use. Microsoft’s electricity consumption has tripled since 2020, contributing to a 30% rise in carbon emissions. Google’s greenhouse gas emissions are up 48% since 2019, with its 2024 Environmental Report citing AI and data centre expansion as major factors.
AI as both problem and solution
Despite the strain, Big Tech leaders argue AI is also a powerful sustainability tool.
Google Chief Sustainability Officer Kate Brandt says: “Our AI-powered efficiency recommendation system for data centres led to a 40% reduction in the energy we use for cooling. There’s a huge opportunity for us to reduce energy consumption across our own data centre fleet and also to make [the technology] more widely available [for others to use].”
Google reports its facilities are now 1.8 times as energy efficient as a typical enterprise data centre. However, its emissions still rose 13% year-on-year in 2024, showing the scale of the challenge. “Scaling AI and using it to accelerate climate action is as crucial as addressing its environmental impact,” says Kate.
Amazon is applying AI to sustainability through Amazon Web Services (AWS), which it claims is up to 4.1 times more energy efficient than on-premises infrastructure. AWS can cut a workload’s carbon footprint by up to 99%, according to the company.
Kara Hurst, Amazon’s Chief Sustainability Officer, explains: “AI and ML can help us meet our climate goals at the speed, scale and urgency our planet requires.”
Amazon’s AI tools are being used to optimise packaging, prevent waste, monitor produce, measure carbon footprints and even prevent deforestation by democratising environmental data.
- Reducing packaging use
- Identifying damaged items to prevent waste
- Monitoring produce to reduce food waste
- Reducing returns by helping customers find the perfect fit
- Measuring the carbon footprint for products
- Preventing deforestation by democratising data
- Using AWS chips to power AI more efficiently
Meta’s renewable and nuclear approach
Meta’s AI growth also comes with higher energy demands. The company has matched 100% of its data centre and office electricity use with renewable energy since 2020.
Its 2024 Sustainability Report outlines plans to explore nuclear power to meet AI’s rising electricity requirements, targeting between 1GW and 4GW of new nuclear capacity in the US by the early 2030s.
“We’re optimistic about the role technology will play in helping us address the climate crisis,” says CEO Mark Zuckerberg. “But the task ahead is massive… the possibilities that our technology will unlock for people only matter if we have a safe and thriving planet.”
Efficiency gains and technical innovation
To reduce AI’s data centre footprint, companies are developing custom chips such as Amazon’s AWS Inferentia and Google’s Tensor Processing Units (TPUs) to improve efficiency per computation. Advanced cooling – including liquid systems and heat reuse – is being deployed to handle the intense thermal load from AI servers.
These technical measures aim to ensure AI’s benefits are not outweighed by its environmental cost. PwC modelling suggests that if AI-enabled efficiencies are widely adopted, they could offset the added power demand from data centres. That could make AI’s net impact on energy use neutral or even positive, optimising power grids, predicting demand, streamlining logistics and cutting waste.
The sector’s challenge is clear: the same AI capabilities that threaten to derail net zero timelines also offer the tools to accelerate them.
Success will depend on how quickly and transparently these innovations are deployed – and whether the sustainability progress can keep pace with the exponential growth of AI infrastructure.
