<p>The numbers paint a stark picture. Global data centre electricity use could more than double by 2030, water consumption may rise from 4.5 trillion litres to over 9 trillion, and carbon emissions could double. These are not distant projections. They reflect decisions being made now about where AI infrastructure gets built, how it gets powered, and what efficiency standards apply.</p><p>For UK businesses, this matters in several ways. Data centres underpin the digital services your business depends on. They also sit inside the supply chains you report on, the energy markets you buy from, and increasingly, the climate commitments you make to customers and investors. As regulators start requiring transparency from the facilities that run cloud platforms and AI tools, those impacts will become harder to ignore.</p><p>What has changed is the speed and scale. AI workloads are pushing data centre energy demand up faster than efficiency improvements can offset it. Consequently, governments are introducing new disclosure rules, grid operators are exploring demand response schemes, and some operators are testing whether waste heat and flexible loads can turn a climate problem into a resource.</p><h2>European disclosure rules target energy and water use</h2><p>The European Commission has proposed mandatory disclosure requirements for data centres. Facilities will need to publish how efficiently they use energy and water. The proposal includes a rating scheme designed to steer investment toward more efficient operations.</p><p>This reflects growing concern that AI infrastructure could strain power grids and water supplies across Europe. It also signals a shift from voluntary reporting to enforceable standards. The framework is still under development, but the direction is clear. Operators will face pressure to measure, report, and improve resource efficiency.</p><p>For businesses using cloud services or colocation, this means greater visibility into the environmental performance of the facilities hosting your data. It also means that procurement decisions may soon include energy and water metrics alongside cost and uptime. Public sector suppliers should expect similar criteria to emerge in UK tender requirements, particularly as net zero commitments tighten.</p><p>The European proposals follow warnings from the United Nations that AI could sharply increase electricity and water demand by the end of the decade. According to UN researchers, global data centre power use reached approximately 448 terawatt hours in 2025. By 2030, that figure could climb to 945 TWh, with AI accounting for 40% of total consumption.</p><p>Water use is projected to follow a similar trajectory. Estimates suggest consumption could exceed 9 trillion litres annually by 2030, driven largely by cooling systems in high-density AI facilities. Carbon emissions from data centres could rise from 189 million tonnes in 2025 to 399 million tonnes by 2030 if current growth trends continue without mitigation.</p><h2>Grid pressure and demand flexibility in the United States</h2><p>In the United States, data centre electricity demand could rise from around 177 to 192 TWh in 2024 to as much as 793 TWh by 2030, depending on AI adoption rates and efficiency gains. Those figures come from analysis by the Electric Power Research Institute and Duke University researchers, who also examined how flexible load management could reduce grid strain.</p><p>Demand response is emerging as one practical answer. This involves data centres temporarily reducing or shifting electricity use during periods of peak grid stress. Facilities surveyed by the Electric Power Research Institute reported peak power reduction potential of between 10% and 30%.</p><p>Duke University researchers estimated that widespread demand flexibility could save between $40 billion and $150 billion in grid infrastructure investment over the next decade. However, this depends on operators designing facilities with flexible loads in mind and grid operators creating markets that reward that flexibility.</p><p>For UK businesses, the relevance lies in supply chain resilience and energy procurement. If your operations depend on cloud services hosted in regions with constrained grid capacity, demand response schemes could affect availability during peak periods. Similarly, businesses with high compute needs may find opportunities to negotiate lower energy costs by shifting workloads to off-peak hours.</p><h2>Cooling design and operational efficiency improvements</h2><p>Cooling accounts for a significant share of data centre energy use. As a result, operators are rethinking how facilities are designed and where they are located. Amazon Web Services now uses air cooling 90% of the year globally, reducing reliance on water-intensive systems.</p><p>This approach depends on climate and facility design. Air cooling works best in cooler regions or when combined with heat recovery systems. In warmer climates, hybrid systems that switch between air and water cooling may offer the best balance between efficiency and reliability.</p><p>Some operators are also exploring ways to reuse waste heat. One example involves using heat from data centres to support carbon capture systems or to produce clean water. At commercial scale, these systems could capture thousands of tonnes of CO2 annually while generating significant volumes of fresh water.</p><p>Software optimisation is another lever. Improved algorithms can increase compute efficiency without expanding energy use. In one reported case, software changes increased AI inference by 33% on existing infrastructure without additional power draw. These gains matter because they allow operators to handle growing workloads without proportional increases in energy consumption.</p><p>For businesses, this translates to two considerations. First, providers that invest in cooling efficiency and software optimisation may offer better long-term cost stability as energy prices rise. Second, if you are reporting Scope 3 emissions from cloud services, ask providers for data on power usage effectiveness and renewable energy procurement.</p><h2>UN calls for transparency and renewable energy commitments</h2><p>UN Secretary-General António Guterres has called on AI firms to disclose their environmental impacts publicly. Specifically, he urged companies to measure and report water, carbon, and land use impacts. He also called for commitments to power data centres with renewable energy by 2030.</p><p>This statement reflects broader pressure on the technology sector to align growth with climate goals. It also acknowledges that current disclosure practices are inconsistent and often lack the detail needed for meaningful comparison or accountability.</p><p>For UK SMEs, the implications depend on your relationship with technology providers. If your business relies on AI tools or cloud platforms, you may face questions from clients, investors, or procurement teams about the carbon footprint of those services. Providers that commit to renewable energy and publish detailed environmental data will be easier to justify in tenders and sustainability reports.</p><p>Furthermore, businesses pursuing Science Based Targets or similar frameworks will need to account for emissions from purchased cloud services. This makes transparency from providers essential. Without it, you cannot accurately measure or reduce your digital carbon footprint.</p><h2>Core facts about data centre resource demand</h2><ul><li>Global data centre electricity use was estimated at 448 terawatt hours in 2025 and could reach 945 TWh by 2030.</li><li>AI workloads are projected to account for 40% of data centre power demand by 2030.</li><li>Water consumption by data centres was estimated at 4.5 trillion litres in 2025 and could rise to 9.3 trillion litres by 2030.</li><li>Carbon emissions from data centres were estimated at 189 million tonnes in 2025 and could reach 399 million tonnes by 2030.</li><li>In the United States, data centre electricity demand could rise from approximately 177 to 192 TWh in 2024 to as much as 793 TWh by 2030.</li><li>Surveyed data centres reported peak power reduction potential of 10% to 30% through demand response schemes.</li><li>Demand flexibility could save between $40 billion and $150 billion in grid infrastructure investment over the next decade.</li></ul><h2>What UK businesses should consider now</h2><p>The first step is understanding where data centres sit in your supply chain. If you use cloud services, software as a service platforms, or hosted AI tools, those services depend on facilities that consume energy and water. Those impacts appear in your Scope 3 emissions, whether you measure them or not.</p><p>Ask your providers for data. Specifically, request information on power usage effectiveness, renewable energy procurement, and water use per unit of compute. Providers that publish this data are easier to work with when you need to report emissions or respond to client sustainability questionnaires.</p><p>Consider how procurement criteria will evolve. Public sector suppliers already face carbon reduction requirements through measures like Procurement Policy Note 06/21. Private sector clients are increasingly asking similar questions. If your competitors can demonstrate lower digital emissions, that may influence contract decisions.</p><p>For businesses with significant compute needs, explore whether your workloads can be scheduled flexibly. Shifting processing to off-peak hours can reduce energy costs and grid impact. Some cloud providers are starting to offer pricing models that reward flexible demand, particularly in regions with high renewable energy penetration during certain hours.</p><p>If your business operates its own server infrastructure, review cooling systems and energy procurement. Even small efficiency improvements can reduce operating costs and carbon footprint. Our <a href="https://sbs.eco/compliance/">ESG compliance support</a> includes carbon footprint assessments that cover digital infrastructure alongside other emission sources.</p><p>Finally, stay alert to regulatory developments. The European disclosure proposals may influence UK policy, particularly if businesses operating in both markets face different reporting standards. Consistency in measurement and disclosure will make compliance easier and more cost effective.</p><h2>Policy uncertainty and investment decisions</h2><p>The speed of AI adoption has outpaced the policy frameworks designed to manage its environmental impact. Consequently, businesses face uncertainty about what standards will apply, when they will take effect, and how enforcement will work.</p><p>This uncertainty affects investment decisions. Operators planning new facilities must weigh efficiency investments against future regulatory requirements that may not yet be finalised. Businesses choosing cloud providers must assess whether today's sustainability claims will hold up under tomorrow's disclosure rules.</p><p>Nevertheless, the direction is clear. Transparency will increase, efficiency standards will tighten, and renewable energy procurement will become a competitive differentiator. Businesses that anticipate these trends will find it easier to adapt than those that wait for mandates.</p><p>For support with carbon reporting, Scope 3 emissions measurement, or preparing for sustainability requirements in procurement, our <a href="https://sbs.eco/net-zero-program/">net zero programme</a> provides practical guidance tailored to UK SMEs.</p><h2>Where to find authoritative guidance</h2><p>The UK government publishes guidance on environmental reporting and energy management through the Department for Energy Security and Net Zero. Their resources include frameworks for measuring and reporting greenhouse gas emissions, which apply to digital infrastructure as part of broader Scope 3 calculations. You can access these at <a href="https://www.gov.uk/government/organisations/department-for-energy-security-and-net-zero">www.gov.uk/government/organisations/department-for-energy-security-and-net-zero</a>.</p><p>For businesses subject to the Streamlined Energy and Carbon Reporting framework, detailed guidance is available at <a href="https://www.gov.uk/government/publications/streamlined-energy-and-carbon-reporting-guidance-for-publication">www.gov.uk/government/publications/streamlined-energy-and-carbon-reporting-guidance-for-publication</a>. This includes how to account for emissions from purchased electricity and cloud services.</p><p>The Institution of Environmental Management and Assessment provides sector-neutral guidance on carbon footprinting and environmental management systems. Their resources can help businesses develop consistent approaches to measuring digital emissions. Further information is available at <a href="https://www.iema.net">www.iema.net</a>.</p><p>Businesses looking for training on carbon measurement and reporting can explore options through our <a href="https://sbs.eco/sbs-academy/">SBS Academy</a>, which covers Scope 3 emissions and supply chain carbon accounting.</p>
