Can the AI Revolution be a Green One?

Artificial intelligence is accelerating at pace and unlocking extraordinary leaps in productivity. It increases data-processing capacity, recognises complex usage patterns, and models carbon pathways with a level of precision that was previously out of reach. But alongside these capabilities comes a growing environmental challenge: AI has a rapidly expanding energy, water, and carbon footprint.

As organisations integrate AI into everyday operations, sustainability can no longer be an afterthought. If the future of business is AI, then AI must be sustainable.

AI’s Rising Environmental Footprint

AI models come with significant energy demands. Depending on the scale, training a model can consume electricity comparable to the annual use of 100 UK homes. However, today most emissions do not come from training—they come from using the models. Every query sent to a large language model requires computation. Each inference carries a carbon cost, and with millions or billions of inferences every day, continuous AI use becomes a major source of emissions.

Data centres already account for an estimated 2–3% of global electricity consumption, and this figure is rising sharply. Google recently reported a 51% rise in emissions from 2019–2024 driven largely by AI expansion, with independent assessments placing the increase closer to 65%. Other major companies—Microsoft, Meta, and Amazon—have also delayed, softened, or revised their net-zero targets as AI demand accelerates faster than renewable energy deployment.

AI is not only energy-intensive; it is resource-intensive. Data centres consume vast volumes of water for cooling, generate waste heat, demand large land footprints, and rely on hardware containing rare minerals extracted through carbon-intensive mining supply chains.

1. Embodied Carbon
Servers, chips, batteries, cooling infrastructure, and the buildings themselves all carry substantial embodied carbon. With hardware cycles as short as 3–5 years, embodied emissions recur far more frequently than in most building typologies.

2. Operational Carbon
This is where the majority of AI’s impact lies. Continuous inference, cooling operations, and uninterruptible power systems all contribute to rising operational emissions—especially as models become more capable and usage volumes expand.

2. Operational Carbon
This is where the majority of AI’s impact lies. Continuous inference, cooling operations, and uninterruptible power systems all contribute to rising operational emissions—especially as models become more capable and usage volumes expand.

Declaration of National Energy Emergency: President Trump declared a national energy emergency, allowing for the suspension of environmental regulations and expediting fossil fuel extraction projects.

Halting Offshore Wind Projects: An executive order was issued to halt leasing and permitting for offshore wind energy projects, impacting the development of renewable energy infrastructure.

Rollback of Environmental Regulations: The administration has initiated rollbacks of various environmental regulations, including those related to air pollution and emissions standards.

A Delayed but Not Defeated Climate Agenda

Trump’s actions undeniably delayed progress, especially at a time when climate science has underscored the urgency for rapid change. However, the sheer scale of backlash from scientists, activists, cities, and businesses has created a new wave of climate leadership—from the bottom up. Many U.S. states, corporations, and municipalities maintained or even increased their climate commitments in defiance of federal policy.

A More Resilient Environmental Movement

The withdrawal era forced environmental and sustainability advocates to rethink strategy. Rather than relying solely on federal leadership, we’ve seen:

• Localised green innovation.
• Shareholder activism demanding sustainable corporate practices.
• Increased investment in ESG (Environmental, Social, and Governance) strategies.

Ironically, Trump’s withdrawals may have helped galvanise a broader coalition of unlikely climate allies – including financial institutions and conservative-leaning municipalities – that now view sustainability not as ideology, but as risk management.

The Risk of Policy Whiplash

One danger of Trump’s rollbacks is the precedent they set: that climate policy in the U.S. can swing dramatically with each administration. This instability makes long-term planning difficult – for governments, businesses, and investors alike. If Trump or a similarly minded candidate wins future elections, the U.S. could again face a regression in sustainability leadership.

A Call for Hard Law and Binding Agreements

Another possibility? Trump’s legacy may encourage climate advocates to push for harder legal protections – turning voluntary or executive-led initiatives into binding legislation that can’t easily be undone. For example, codifying emissions targets into law or creating bipartisan-backed green infrastructure bills could insulate climate progress from political whims.

Trump’s withdrawals from sustainability initiatives were more than symbolic, they altered the trajectory of U.S. climate leadership at a pivotal moment. But they also exposed the fragility of progress built on executive orders and public goodwill.

The possibilities now are twofold. On the one hand, we risk continued instability in environmental policy. On the other, we have a unique opportunity to build a more grounded, resilient, and bipartisan foundation for sustainability – one that endures beyond political cycles.