Expert Insight: The Advent of AI, Cloud Infrastructure, and the Impact on Carbon Emissions for Business

Expert Insight: The Advent of AI, Cloud Infrastructure, and the Impact on Carbon Emissions for Business

What is Cloud Infrastructure (CI) and Artificial Intelligence (AI), and Why Do They Matter?

Since the advent of the internet and its integration into daily business operations, Cloud Infrastructure (CI) has become a cornerstone of modern organisations. Increasingly, businesses rely on CI systems - often delivered as Software as a Service (SaaS) - to provide critical functions, including data storage, accounting, customer relationship management (CRM), and videoconferencing.

As these systems scale in complexity and capacity, their physical infrastructure - data centres, servers, and processing facilities - expands accordingly, increasing demand for energy and carbon-intensive materials.

Now, with the proliferation of Artificial Intelligence (AI) tools, governments and industries are embracing the potential economic and social benefits of digital solutions. From automating mundane tasks to advanced data analysis, AI promises to revolutionise how businesses operate.

However, there is a hidden cost: AI’s growing energy demands and reliance on ever-expanding CI infrastructure amplify its environmental footprint. As businesses accelerate digital adoption, addressing the carbon emissions from these systems must become a priority.

 

How Do AI and CI’s Energy Needs Shape Climate Action?

The United Nations has underscored the environmental implications of AI and CI in its Sustainable Development Goals (SDGs). These technologies contribute significantly to:

·         SDG 9: Industry, Innovation, and Infrastructure

·         SDG 12: Responsible Consumption and Production

The 2024 UN Digital Economy Report revealed that, as of 2020 – prior to the widespread accessibility of AI - the Information and Communications Technology (ICT) sector accounted for 1.5–3.2% of global emissions. This figure has likely grown significantly due to the increased scale of AI training and cloud infrastructure development.

At Carbon Neutral Britain, we recognise that as organisations adopt CI and AI solutions, their energy consumption and emissions are likely to rise. This is particularly likely for businesses operating their own data centres, which consume significant electricity and require substantial carbon-intensive resources to construct and maintain.

Industry experts anticipate that reporting frameworks such as the GHG Protocol and standards like ISO 14064 may eventually integrate specific requirements for tracking the emissions of AI and CI usage as part of Scope 3. This shift reflects a growing need to better account for the climate impacts of these rapidly expanding technologies.

 

Current Discourse on CI and AI Emissions

Discussions on the environmental impact of CI and AI are still in their infancy. However, the financial services industry is emerging as a leader in advocating for greater emissions transparency, particularly from an investment risk perspective.

Organisations such as the CFA Institute stress the importance of understanding the emissions associated with scaling and training AI models. These emissions are not trivial. For instance, training GPT-3, a prominent AI model, consumed approximately 1,287 MwH of electricity, equivalent to around 502 tCO₂e. Other models exhibit similarly significant emissions, highlighting the need for mitigation as AI becomes further embedded in SaaS platforms.

As carbon reporting evolves, businesses must adapt to new requirements that demand more granular data on their emissions, including those linked to CI and AI. This evolution will reshape how organisations assess and manage their environmental impact.

At CNB, we ensure that all calculations and reporting aligned with the principles and guidelines of the GHG Protocol, ISO 14064, and the Science Based Targets (SBTs). These standards provide the foundation for reliable and consistent emissions accounting, ensuring organizations are well-prepared for future reporting obligations, and are the most likely to expand reporting requirements to include the use of CI, SaaS, and AI.

 

How to Integrate CI and AI with Sustainability in Mind

For larger companies that operate their own CI systems, integrating sustainability considerations into these systems may be more straightforward, as a result of more direct oversight into energy usage and carbon intensity.

For small and medium-sized enterprises (SMEs), however, the path forward can be more complex. SMEs often rely on external service providers and may lack visibility into their emissions. Nevertheless, organisations can align CI and AI usage with sustainability goals in a number of ways.

Where primary data is unavailable, businesses may look to engage with suppliers to request emissions information, fostering transparency across the value chain.

Collaborating with suppliers who report their emissions, and implementing review processes that prioritise environmental performance can further drive sustainability efforts.

While reporting emissions from CI and AI is not yet mandatory, organisations should proactively collect and document related data, including software types and their emission factors, to prepare for potential future compliance requirements.

Optimising infrastructure is another critical step. Ultimately, the most effective way to reduce emissions is to change and reduce usage. While many may look to renewable energy in the form of Renewable Guarantees of Origin (REGO) and Power Purchase Agreements (PPA), such renewable energy is sent directly into the electricity grid, rather than directly to the customer, and so do not directly reduce the emissions of the end user.

Conclusion

As businesses increasingly rely on CI and AI integrated SaaS, they must also address the associated environmental costs. Organisations that take proactive steps to measure, manage, and mitigate these emissions will not only align with emerging standards but also position themselves as leaders in the global effort to combat climate change.

At CNB, we empower organisations to both measure and address their emissions, as a key step in their pathway to Carbon Neutrality.

One of our key deliverables, through our tailored Carbon Reduction Plans (CRPs), is to provide organisations with a clearer path toward both identifying and reducing future emissions. As part of this, we actively highlight to organisations the need to better prepare for future reporting periods by putting systems in place to efficiently and accurately identify relevant data year-round, with energy consumption playing an increasingly large role in this as demand increases.

And, while not currently a reporting requirement, all organisations should consider paying attention to the various types of CI-reliant software and services being used – including AI – and, where possible, implement supplier selection-and-review processes which consider the environmental credentials of such suppliers.

 

Expert Insight Written by:

Peter Westbury - BSc (Hons – Climate Finance and Decarbonisation), ESG Analyst, ISO 14064 and GHG Protocol Specialist - Environmental Consultant at Carbon Neutral Britain.

 

References:

https://blogs.cfainstitute.org/investor/2024/10/31/the-hidden-environmental-costs-of-tech-giants-ai-investments/?preview_id=110519&_thumbnail_id=110525#_ftn12

https://sciencebasedtargets.org/blog/net-zero-jargon-buster-a-guide-to-common-terms#carbon-neutral

https://www.statkraft.com/newsroom/explained/guarantees-of-origin-ensuring-100-per-cent-renewable-power-in-europe/#:~:text=A%20Guarantee%20of%20Origin%20tracks,and%20where%20is%20it%20located%3F

https://www.technologyreview.com/2022/11/14/1063192/were-getting-a-better-idea-of-ais-true-carbon-footprint/

Luccioni, A.S., Viguier, S. and Ligozat, A.L., 2023. Estimating the carbon footprint of bloom, a 176b parameter language model. Journal of Machine Learning Research, 24(253), pp.1-15.