Skip to main content
Marsh logo

Article

The Sustainability Challenge: Balancing Data Centre Growth with Environmental Responsibility

This article examines the environmental impact of energy and water consumption, the push for net-zero emissions, and innovative solutions in the tech industry.

11/07/2025 · 4 minute read

As the data centre industry navigates the complexities of balancing growth with environmental responsibility, prioritising sustainable energy and resource management is essential. Contact our experts to discuss how they can help support your business's future development.

AI is rapidly transforming the world, but this progress could potentially be at a cost. While AI offers remarkable opportunities, the full environmental impact of the rapid rollout of data centres to meet this demand is unknown. Could AI help us tackle the climate emergency or exacerbate the issue?

Energy consumption

The energy consumption associated with AI is likely to rise as usage increases. For example, a large hyperscale data centre consumes as much electricity annually as 400,000 electric cars. Much of this energy still comes from traditional energy sources that emit greenhouse gases, making a data centre’s carbon footprint a growing issue.

Several tech firms have set ambitious goals to achieve net zero carbon emissions by 2030 or 2040, and several are among the world’s biggest corporate buyers of clean energy. Additionally, fund managers and banks are under increasing pressure to decarbonise their portfolios, prompting them to scrutinise potential data centre investments.

This scrutiny includes a data centre’s secondary and backup energy supplies. While a data centre’s main source of power could be renewable energy, it may rely on diesel-powered generators in the event of a power failure.

Data centres could spark debates over energy prioritisation as national energy supplies come under strain. The power used by data centres in the UK is set to increase six-fold in the next ten years, according to the National Grid. In Ireland, data centres could account for over a third of the country’s electricity by 2026. As a result, there is an effective moratorium on the construction of new data centres in or around Dublin until at least 2028.

Water consumption

Some data centres also consume vast quantities of water to cool the banks of servers used to process information requests. Global AI demand could consume up to 6.6 billion cubic metres of water by 2027, equivalent to half the water used by the UK. With a quarter of the world’s population lacking access to safe drinking water and sanitation, this issue could become increasingly pressing for politicians, advocacy groups, and the public.

As a result, data centres are likely to face more stringent regulations. Germany has passed an Energy Efficiency Act requiring data centres to reduce or reuse waste heat, boost their power use effectiveness, and eliminate power from traditional energy sources by 2027. Other governments are likely to follow Germany’s lead, while the United Nations’ Environmental Programme (UNEP) recommends standardised procedures for measuring the environmental impact of AI and is urging governments to require companies to disclose the environmental impact of their AI products and services.

A positive outlook

However, it’s not all negative news regarding AI and the environment. Tech companies are already looking to make their algorithms and models more energy efficient and use less resources. New photonic chips are being tested that can offer a faster and more energy-efficient option for the most complex and demanding AI applications, such as high-speed telecommunications. By using light to transmit and process data, photonic technology is expected to transform AI, opening the door to high-speed processors that are capable of learning in real-time.

AI tools are being used to tackle climate change. For example, by identifying complex patterns, processing huge data sets, and optimising resource management. It is also being used to improve the efficiency of energy grids, decrease fuel use in shipping, and detect methane gas leaks.

Conclusion

Embracing environmental stewardship in AI operations is not just a regulatory obligation, it is also a strategic imperative that can enhance reputations, foster stakeholder trust, and drive long-term success.

Marsh’s Data Centre Insurance and Risk Management Services team is dedicated to providing comprehensive risk management and insurance for clients throughout the entire lifecycle of data centre operations, from build, to operations, and owning data centres. For more information please contact us to discuss how we can help support your business's future development.

Our people

Amy Barnes

Amy Barnes

Global Head of Energy & Power, Marsh

  • United Kingdom

Jason Payne

Jason Payne

Data Centre Lead, Real Estate Practice, Marsh UK

  • United Kingdom

Related insights

Innovative renewable energy network displayed in high-tech visualization, emphasizing sustainability and connectivity.

Article

Regulatory Challenges and Sustainability in Data Centres

11/07/2025
Construction engineers discussion with architects at construction site or building site of highrise building with Surveying for making contour plans is a graphical representation of the lay in land.

Article

Ensuring resilience: The role of insurance in data centre construction

23/06/2025
Flow of digital information

Article

Powering the future of data centres

03/03/2025
Placeholder Image

Article

Why risk mitigation starts with contractual negotiations

03/03/2025