AI and Sustainability: CBRE’s Data Centre Shift

Data centres play a central role in the global digital infrastructure, hosting the cloud platforms, applications and AI models that drive today’s technology landscape.

But their high energy demand has led many to question whether their environmental footprint can align with global sustainability goals.

CBRE, the global real estate and infrastructure services company, challenges that view.

It argues that data centres can become part of the sustainability solution, not just the problem, through intelligent deployment of AI, better integration with renewable power and strategic use of heat recovery.

In its Data Centers, AI and Sustainability: Navigating the Carbon Paradox report, CBRE outlines how technology and sustainability objectives can converge, even in high-energy sectors.

The rising energy cost of AI

Data centres are critical for cloud computing, ecommerce, media streaming and especially AI.

Yet, they are also resource-intensive, consuming large amounts of electricity and water.

In 2024, they are estimated to use between 415TWh and 460TWh — about 2% of global electricity consumption.

Projections indicate that by 2030, this figure could nearly double, largely driven by AI-related computing demands, with upper estimates nearing 1,065TWh — comparable to the entire annual energy use of Japan.

The growth of AI itself is a key factor in these projections.

According to Digiconomist, by 2027, AI is expected to require between 85TWh and 134TWh each year.

That figure equates to Singapore’s energy consumption over 18 months to three years.

These forecasts highlight a growing tension between the benefits of AI technologies and their energy and carbon costs.

However, CBRE suggests this narrative oversimplifies the issue.

While AI contributes to the problem, it also offers tools to manage and reduce energy usage.

“The increased resource demand associated with AI may contribute to the perception that AI and data centres are at odds with sustainability goals,” CBRE says.

However, it is important to recognise the significant energy and carbon emissions reduction benefits that this technology can provide.”

Intelligent optimisation and grid integration

CBRE identifies three strategies that enable data centres to reduce their overall energy consumption while still meeting demand for advanced AI workloads. At the centre of these strategies is the use of AI for data-led optimisation.

Facility electricity consumption optimisation is the first of these.

By using AI to analyse real-time data on energy use, occupancy and external factors such as weather, data centres can align operations with periods of lower grid emissions and greater renewable availability.

This also reduces wear on equipment, as AI models minimise unnecessary start-stop cycles, improving both efficiency and equipment lifespan.

The second approach focuses on smart grid integration.

AI enables data centres to forecast energy supply and demand more accurately, helping them operate in sync with the grid.

Importantly, data centres can move from being simple consumers to prosumers — entities that both use and generate energy.

By sending surplus electricity or repurposed waste heat back into the grid, data centres help balance loads, reduce peak demands and contribute to wider decarbonisation efforts.

CBRE highlights the opportunity to convert waste heat into a resource.

Instead of letting heat escape, data centres can transfer it into nearby district heating networks, especially in colder climates, cutting both heating emissions and overall facility waste.

Strategic siting and heat reuse

Location matters in the effort to make data centres more sustainable.

The third strategy outlined by CBRE is strategic colocation and heat reuse. AI workloads, such as training large models or performing batch inference, are often tolerant of higher network latency.

This allows data centres to be built in areas that have lower-carbon electricity and existing infrastructure for energy reuse.

A prime example is Sweden.

Its data centres can run on hydropower and feed into district heating systems.

CBRE estimates that a Swedish facility using this model emits up to 40 times less CO₂ than a similar centre based in Singapore, where the grid is more carbon-intensive and the climate hinders efficient cooling.

CBRE’s Chief Sustainability Officer Robert Bernard explains: “The potential return is remarkable.

“Our analysis shows that these strategies can deliver massive energy and carbon savings that far outweigh the resources invested. 

“This isn’t just good for our planet — it’s smart business. 

“When we design these data centres thoughtfully and intentionally, these facilities can become active participants in the green economy.”

By reframing data centres not simply as energy consumers but as platforms for AI-driven optimisation and sustainable integration, CBRE presents a model in which technology infrastructure aligns with environmental priorities.

AI is not only consuming power — it is also redefining how efficiently that power can be used.

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today. 

AI Magazine is a BizClik brand