Page Contents
Related areas
Key contacts
For many organisations, AI is viewed as a strategic priority. At the same time, businesses are facing increasing expectations from regulators, investors, customers and other stakeholders to demonstrate meaningful progress on sustainability objectives. While both agendas are often pursued simultaneously, AI and sustainability initiatives are generally viewed as completely separate workstreams. It is increasingly important for Boards and senior management teams to understanding how the AI and sustainability strategies interact.
Globally data centres accounted for 1.5% of worldwide electricity consumption[1] in 2025, however in Ireland the share of total metered electricity consumption attributable to data centres reached 23%[2]. Although AI represents only one category of data centre activity, the rapid growth in AI adoption has focused attention on the environmental impacts associated with digital infrastructure.
While AI can deliver significant operational and sustainability benefits, it also relies on a complex physical infrastructure of data centres, computing power, energy and natural resources. This raises important governance questions such as:
• are the AI and sustainability strategies aligned?
• have the environmental impacts associated with AI deployment been considered as part of decarbonisation efforts?
• should sustainability considerations play a greater role in the procurement and implementation of AI tools?
In this article, we examine the interaction between AI and corporate sustainability strategies. We also highlight some key practical considerations for Boards and senior management teams when implementing and overseeing both strategies.
AI adoption and climate transition plans
It is important to recognise that while AI is only one type of work that is performed by data centres, it is particularly energy intensive. According to the International Energy Agency (IEA)’s Energy and AI report[3], while traditional data centres use between 10 and 25 MW of power, demand from hyperscale AI centres can exceed 100MW or the equivalent of the annual electricity consumption of 100,000 homes. According to 2025 data centre electricity consumption data published by the Central Statistics Office between 2015 and 2025 data centre consumption grew by 584% from 291 GwH in the first quarter of 2015 to 1,991 GwH in the fourth quarter of 2025[4].
While this is a highly energy intensive physical ecosystem, the majority of organisations are not building AI infrastructure, instead they are buying AI products and services typically focused on operational efficiencies. As AI infrastructure becomes more sophisticated, the energy usage for individual queries will continue to reduce with the IEA noting that “Simple text queries now typically consume less electricity than running a television over the same period of time”[5]”. New AI applications continue to be launched that consume significantly more energy than a simple text query highlighting the importance of understanding the use cases for which businesses intend to adopt AI tools and processes.
Identifying how AI will be used will assist with determining the impact such usage could have on the company’s ability to successfully implement its sustainability strategy. From an energy consumption perspective, there will likely be a significant difference between providing employees with access to AI tools to assist with their day to day work and introducing AI tools into key business processes to create efficiencies for overall operations. Either way, it is important for organisations to understand the environmental impacts associated with the technologies they deploy.
To understand the potential impacts that increased AI usage could have on meeting sustainability targets, organisations should consider whether increased AI usage could have a material impact on scope 2 or scope 3 emissions, emissions reduction targets or broader sustainability objectives. Such impacts could include, on the one hand, increased electricity consumption and on the other, energy efficiencies achieved through the use of AI tools. If a climate transition plan has been adopted, it is worth reviewing this plan to determine whether AI usage has already been factored in or whether any updates might be needed.
One of the challenges currently facing businesses is that most providers of AI products are not sharing granular energy or water consumption data at the level of individual queries or products with the businesses using these products. This makes accurate calculations by individual businesses difficult. Some providers do publish estimates, for example, Google measures the energy, emissions and water impacts relating to Gemini prompts[6]. Third party calculations may also be helpful in terms of estimating use. Similar issues exist in relation to other types of sustainability information creating challenges for those businesses reporting sustainability information. Transparency relating to the availability and reliability of this data is important, particularly where organisations are making claims regarding the sustainability benefits of AI enabled products, services or initiatives.
It is useful for organisations to consider how AI can support sustainability objectives. AI tools are being used in ever-increasing ways when it comes to sustainability including to assist with emissions monitoring, energy optimisation, supply chain mapping and to create reporting efficiencies. Understanding both the environmental costs and the sustainability benefits associated with using AI can help businesses to align their AI and sustainability strategies and support informed decision-making.
Sustainable procurement of AI
AI is a product of complex supply chains. Understanding the links in the chain can enable businesses to request the most pertinent information when conducting due diligence as part of the procurement process. In addition, increased focus on sustainability due diligence means it may be useful for organisations to consider the environmental and human rights impacts across the entire AI value chain. For those within scope of the Corporate Sustainability Due Diligence Directive, consideration may need to be given to whether any environmental or human rights impacts associated with this particular chain of activities constitute adverse impacts requiring further assessment.
In this context, it is useful to consider the environmental impacts of each of the stages of the AI model lifecycle:
• the AI infrastructure. Certain critical and rare earth materials are required for the key components of data centres including computer chips. Cooling systems that use large amounts of water are required to keep servers at optimal temperatures. The energy and resources used in the data centre construction process also needs to be factored in.
• training models. Energy consumption for training varies substantially depending on model size, complexity and hardwire configuration. However, it is recognised that training is an energy intensive process.
• model use. As set out above, the lack of data makes it difficult to estimate energy consumption for this stage. Assessments conducted by the IEA show that using generative AI to generate a single short video (6 seconds in length and 8 frames per second) can be as energy intensive as charging a laptop twice[7].
When making procurement decisions on specific AI tools and providers, consideration should be given to:
• the sustainability credentials of the providers including any environmental commitments made. For example, if the provider uses its own data centres the types of questions that could be considered include: what is their energy mix, is renewable energy being used, what type of energy is used for back-up generation, what are their water consumption practices, do they make use or are they planning to make use of waste heat and do they have a climate transition plan in place? For those providers using hosted data centres, ensuring that this information is available.
• the type of data around energy and resource use the providers can share.
• where AI processes have been introduced into key business operations, the operational resilience procedures in place to determine whether any additional information needs to be provided to deal with. For example, any failures in the AI infrastructure or any direct or indirect impacts that grid capacity or other infrastructure constraints could have on service availability.
Consideration should be given to the supply chain for AI products that are already being used and whether sustainability factors were taken into account as part of the procurement process. This is particularly important where AI processes have been introduced into key business operations or may represent a material aspect of the organisation’s digital infrastructure strategy. Documenting the approach taken will ensure that good governance processes are in place.
Implementation of strategy
Responsibility for the AI and sustainability strategies often rest with different functions within an organisation, typically with the IT and sustainability teams respectively. Input from other functions such as legal, compliance, procurement, risk and finance is essential to ensure that all potential impacts, risks and opportunities are explored.
Key questions that Boards and senior management teams should be asking include:
1. Have ESG considerations been taken into account in AI implementation projects?
a. Have appropriate sustainability considerations been incorporated into procurement processes for AI providers?
b. Is there sufficient information to provide to the sustainability team on energy consumption associated with each of the AI tools being used by the business?
2. Has consideration been given to how AI tools could assist in the implementation of the sustainability strategy?
3. Has the impact of the AI strategy on decarbonisation efforts been calculated and documented? If a climate transition plan is in place, are any updates needed?
To date, businesses have been focused on understanding the potential AI use cases. Organisations that take a holistic view of AI’s environmental footprint, focusing not only on energy and emissions but also on factors such as critical mineral extraction and freshwater use, and incorporate these considerations into their sustainability strategy, will be well placed to identify and mitigate emerging sustainability risks while capturing the opportunities presented by AI driven transformation.
For further information in relation to this topic, please contact Jill Shaw, ESG & Sustainability Lead, , Mark Ellis, Partner, Technology or any other member of the ALG ESG & Sustainability team.
Date published: 22 September 2026
[1] https://iea.blob.core.windows.net/assets/3179f7f8-01f6-4dd6-bffa-c9f7b73f1dc9/KeyQuestionsonEnergyandAI.pdf
[2] https://www.cso.ie/en/releasesandpublications/ep/p-dcmec/datacentresmeteredelectricityconsumption2025/keyfindings/
[3] https://iea.blob.core.windows.net/assets/de9dea13-b07d-42c5-a398-d1b3ae17d866/EnergyandAI.pdf
[4] https://www.cso.ie/en/releasesandpublications/ep/p-dcmec/datacentresmeteredelectricityconsumption2025/keyfindings/
[5] https://iea.blob.core.windows.net/assets/3179f7f8-01f6-4dd6-bffa-c9f7b73f1dc9/KeyQuestionsonEnergyandAI.pdf
[6] https://cloud.google.com/blog/products/infrastructure/measuring-the-environmental-impact-of-ai-inference
[7] https://iea.blob.core.windows.net/assets/de9dea13-b07d-42c5-a398-d1b3ae17d866/EnergyandAI.pdf
