Newsletter 3: Levies, Divides, and Data Centres

As governments scramble to balance growth and green goals, the world of AI continues to shape – and be shaped by – urgent sustainability challenges. This week, proposals to tax AI, tensions over data centre footprints, and new global findings on climate inequality reveal that innovation and equity remain deeply entwined.

1. Tax on AI and Crypto Could Fund Climate Action

Laurence Tubiana, a principal architect of the Paris Agreement and now co-lead of the Global Solidarity Levies Task Force, has called for the introduction of levies on artificial intelligence and cryptocurrency to help fund climate mitigation strategies. As she explained in a recent interview with The Guardian, her proposal is rooted in the fact that both sectors are highly energy-intensive – bitcoin mining alone consumes as much electricity as Poland does annually. However, the proposal faces political resistance, particularly in the United States, where Donald Trump has expressed strong support for both industries. Tubiana’s initiative underscores a broader shift toward ensuring that the climate costs of emerging technologies are shared more equitably on a global scale.

2. AI and Climate Action: Opportunity or Digital Divide?

The UNFCCC Technology Executive Committee has published a new technical paper examining how artificial intelligence can accelerate climate action in developing countries (UNFCCC, 2025). Unveiled at the AI for Good Summit in Geneva as part of the #AI4ClimateAction initiative, the paper outlines AI’s potential to reduce emissions, optimize energy systems, and enhance resilience to extreme weather events.

The report also highlights significant structural challenges. Many developing nations in the Global South lack the necessary digital infrastructure and access to high-quality climate data, which limits their ability to benefit from AI-enabled solutions – advantages that are increasingly available to countries in the Global North. This disparity is accentuated by the fact that these vulnerable states are among the most exposed to the impacts of climate change (Christensen, 2024) and therefore have the most to gain from AI-driven adaptation and mitigation tools.

To address this digital divide, the UN recommends targeted investment in digital infrastructure, inclusive and locally-relevant AI design, and open-data platforms that support climate resilience in underserved regions.

3. Clash in Teesside: Data Centre vs Clean Energy

Teesworks Ltd has submitted a planning application for a 500,000 square metre facility that could become Europe’s largest AI data centre (Mailk, 2025, Gross et al., 2025). If approved, the site would anchor the UK’s second AI Growth Zone, placing Teesside and Northeast England—a region historically overlooked in national industrial strategy—at the centre of Britain’s emerging AI infrastructure.

However, the same land is also earmarked for BP’s blue hydrogen and carbon capture project, a core component of Energy Secretary Ed Miliband’s Net Zero strategy. Known as the East Coast Cluster, the initiative aims to be the UK’s first operational carbon capture, usage and storage (CCUS) project—capable of powering up to one million homes with low-carbon energy and enabling the decarbonisation of key industrial sectors (DESNZ, 2024).

This emerging land-use conflict reveals a deeper tension in UK industrial governance: between those prioritising AI-driven competitiveness and those focused on delivering climate commitments. According to reports, a tech department official remarked that “the AI agenda would be way more advanced if it wasn’t for Miliband’s focus on net zero” (Gross et al., 2025). While not all AI advocates oppose environmental goals, the potential approval of the data centre over the CCUS project would suggest that industrial growth is being politically prioritised over sustainability.

4. AI in SDG Research: Potential Undermined by Inequality

In a comprehensive meta-analysis of 792 peer-reviewed studies, Ghor et al. (2025) found that artificial intelligence holds significant potential for advancing sustainable development. Applications include mapping emissions, predicting climate variability, and optimizing resource use across sectors such as agriculture, energy, and water systems.

However, the authors highlight a persistent imbalance: the majority of studies originate from high-income countries and are designed for high-income contexts, limiting the relevance and accessibility of AI solutions for the Global South. They argue that AI systems continue to reflect the values, assumptions, and priorities of the Global North, while researchers, institutions, and communities in the Global South are largely excluded from shaping AI’s design, governance, or deployment.

At the same time, ethical and regulatory frameworks—especially those in emerging markets—are struggling to keep pace with the rapid evolution of AI technologies, raising concerns about inclusivity, fairness, and long-term sustainability (UNESCO, 2021, OECD AI Principles).

5. Meta’s Superclusters: A Manhattan-Sized Trade-off

Mark Zuckerberg has announced that Meta will invest hundreds of billions of dollars to build next-generation AI systems, including a data centre reportedly as large as Manhattan (Sullivan, 2025). The company’s first AI “superclusters” include Project Prometheus (1 GW) in Ohio and Project Hyperion (5 GW) in Louisiana, which are poised to become some of the world’s largest AI-dedicated infrastructure sites (Wheeler, 2025, Zeff, 2025).

While the scale of these facilities raises serious concerns given AI’s heavy demands on water and energy, Meta has outlined a series of sustainability commitments:

• Use 100% renewable energy across all global operations (Meta, 2025)
• Deploy AI to optimize airflow in cooling systems, cutting fan energy use by 20% (Meta, 2024)
• Invest in carbon capture projects near the Louisiana site (Geiger & Thomas, 2025)
• Restore more water than it consumes in high-stress watersheds (Meta, 2025)

Despite these pledges, concerns persist. AI data centres rely on water-intensive cooling systems that may strain local resources. Critics argue that Meta’s mitigation efforts may fall short of offsetting the full ecological footprint.

Final Thought

From Meta’s superclusters to Teesside’s planning war, this fortnight reveals a critical tension: sustainability is being repositioned – not as a guiding principle, but as a secondary concern in the race for AI dominance. Whether through tax resistance, data inequality, or stalled climate infrastructure, these developments show that without meaningful regulation and inclusive governance, AI risks entrenching the very crises it claims to solve.

Thanks for reading
Gus Healey

https://www.linkedin.com/in/gus-healey-aa2888236/

Bibliography

Harvey, F. (2025) … link
UNFCCC (2025) … link
Malik, S. (2025) … link
Gross, T. et al. (2025) … link
DESNZ (2024) … link
Energy Live News (2025) … link
Ghor, C. et al. (2025) … link
Sullivan, H. (2025) … link
Wheeler, K. (2025) … link
Zeff, M. (2025) … link
Meta (2025) … link
Geiger, D. & Thomas, E. (2025) … link