Artificial Intelligences Energy Paradox 2025

Of the ecosystem enablers, regulators and policy- makers in particular are critical to ensuring a sustainable AI future. In considering regulations, both government regulations and industry-led initiatives are crucial, as government rules provide legal frameworks while industry-led initiatives and voluntary actions depend on self-enforcement. Together, they play different but complementary roles in enabling AI. One example is the EU AI Act, which categorizes AI systems by risk, imposing strict requirements on high-risk applications for safety, transparency and accountability, while also cultivating innovation.21 While these regulations aim to drive efficiency and accountability, they may incur compliance costs and unintended consequences. Well-designed incentives, on the other hand, can facilitate continuous improvement and innovation, emphasizing the need for a balanced regulatory approach. Another key consideration is balancing data sovereignty requirements with efforts to locate data centres near clean energy sources. While renewable energy reduces environmental impact, data laws often mandate local storage for privacy and security. This creates tension between the goals of minimizing emissions and meeting regulatory demands, necessitating strategic planning to align both goals. Within infrastructure, regulatory frameworks and policies can support several critical areas including transmission system planning and siting, improving electricity market structures and enabling greater access to carbon-free electricity sources. Customer affordability is also an important area of note, as rate designs are meant to drive fair allocation of costs for customers while ensuring reasonable rates. A key challenge in rate design for data centre growth is balancing the need for scalable, cost-effective electricity pricing with the goal of protecting customer affordability. As demand for AI increases, rates will need to be structured to support large-scale electricity needs without placing undue cost burdens on customers, while also promoting efficiency and sustainability. This is challenging however, and can take many forms.22,23 Additionally, establishing green mandates, aligning with regional emissions targets and improving access to renewable energy are also key steps for sustainability, along with implementing water conservation and energy reduction policies. Together, measures such as these can promote more sustainable AI. 3.4 Regulatory and policy enablers Financial support for sustainable AI can come in several forms (e.g. tax credits or deductions), including incentives for using renewable energy and selecting environmentally friendly sites. Companies like Crusoe are using incentives, such as Bill C-59 in Canada, to support CCS activities.24 As similar financial incentives are made available, frameworks can be expanded to require societal benefits, such as job creation, economic development and community investment. Government investments in infrastructure and financial support for land development and environmental mitigation could further enhance the appeal of data centre locations and facilitate growth. Designing these incentives appropriately with relevant stakeholders can enhance the overall economic impact of data centres and promote a greener future.3.5 Financial incentive enablers Within infrastructure, regulators can promote rate designs that ensure data centres help to upgrade costs while maintaining affordability for customers. –Market development enablers that can create a conducive environment for sustainable AI solutions, encourage collaboration among stakeholders and facilitate adoption of green technologies –Together, these ecosystem enablers could form a robust foundation for advancing sustainable AI. While not explored in this paper, the above could also include ethics and governance considerations. Learn more in the Forum’s 2024 white paper: Governance in the Age of Generative AI: A 360º Approach for Resilient Policy and Regulation.20 Artificial Intelligence’s Energy Paradox: Balancing Challenges and Opportunities 16