July 1: AI for Planet

This week’s look at all things AI & Sustainability – Google’s Problematic Emissions, Rising Water Issues, The Big Beautiful Bill & Climate Positives from AI?

Having written an AI in Trading Newsletter for the last 6 months, we are now expanding to AI & Sustainability. This newsletter is aimed at Heads of Trading Desks navigating an increasingly complex landscape where technological innovation intersects with environmental and regulatory pressures. We aim to highlight how AI’s explosive growth—while offering tools for sustainable finance, climate risk modeling, and supply chain transparency—is simultaneously driving significant increases in energy and water consumption, with mounting scrutiny from regulators and the public. For trading leaders, understanding these trends is crucial not only for risk management and ESG compliance but also for spotting emerging opportunities in green technology, carbon markets, and sustainable investment strategies. Staying informed on AI’s environmental footprint ensures trading desks remain competitive, resilient, but also aligned with clients’ evolving sustainability priorities.

AI & Sustainability: Progress or Problem?

As AI continues to evolve at a breakneck pace, so too do concerns around its ecological footprint. This week’s developments reveal a growing tension between innovation and impact – where AI’s potential to accelerate sustainability risks being undermined by its very infrastructure.

1. Google’s Carbon Emissions Up 51%

In its latest 2025 Environmental Report, Google revealed its ambition-based emissions grew 11% in the last year, marking a 51% increase since 2019. That’s equivalent to emitting 11.5 million metric tons of CO2 – or powering 1.4 million UK homes for a year.

Though the company says it remains committed to their “climate moonshots”, they believe achieving them has become more complex due to a number of external factors – principally the booming AI demand. As The Verge (2025) reports, Google says the rapid pace of AI development has made it difficult to predict future energy use and emissions.

Worryingly, these figures don’t include all supply chain emissions, meaning the full climate cost of AI could be even higher.

This highlights a growing dilemma: AI may help solve climate challenges – but its infrastructure is threatening the very targets it aims to support.

Still, Google offers a silver lining. Today, 45% of its data centres run on 80%+ carbon-free energy, and they now deliver six times more computing power per unit of electricity than 5 years ago. The company argues that while emissions are rising in the short term, AI’s long-term potential to reduce emissions may ultimately tip the balance.

2. AI’s Thirst: A US Sustainability Challenge

While much attention has been paid to AI’s carbon emissions, its water consumption is a growing concern—especially in the U.S., where many data centres are concentrated in drought-prone regions.

According to Business Insider (2025), nearly 40% of U.S. data centres are located in highly water-stressed areas. In Arizona, Microsoft and Meta consume millions of gallons of water daily, putting pressure on local agriculture and the already strained Colorado River, a lifeline for 40 million US residents.

Despite drought legislation like Arizona’s Groundwater Management Act, the state has become a major hub for data infrastructure. Officials in cities like Aurora are warning that unchecked data centre growth may threaten both water supply and urban development. Farmers report losing as much as a third of their arable land due to competition over water.

Big tech companies have responded with pledges to be water positive by 2030, and some new facilities use closed-loop cooling to limit direct use. But questions remain about the timeline and effectiveness of these measures.

3. The ‘One Big Beautiful Bill’ and a 10-Year AI Deregulation Pause

A provision buried in Trump’s sweeping tax bill, the ‘One Big Beautiful Bill Act’, would place a 10-year moratorium on any state or local-level regulation of AI models or systems (The Guardian, 2025).

Whilst proponents of AI state advances in the field will aid the fight against climate change, not everyone is as optimistic. Harvard public health scholar Gianluca Guidi believes that by limiting oversight we would be ignoring the industry’s environmental toll and slowing current progress toward a net-zero future. Similarly, the direct of research at the Distributed AI Research Institute claims such an attitude is a greenwashing manoeuvre in which “Big tech is mortgaging the present for a future that will never come”.

Opposition from Democrats, such as Massachusetts senator Ed Markey, may yet strike the clause from the final bill, but the message is clear: deregulation has become a geopolitical strategy in the race for AI supremacy.

4. Hope on the Horizon? AI as a Climate Solution

A recent Grantham Institute Study argues that if strategically deployed in sectors such as power, transport, and food consumption, AI could cut global emissions by 3.2 to 5.4 billion tons annually by 2035.

Applications include:

· Optimizing energy grids and renewable output

· Nudging sustainable behaviour

· Managing climate adaptation and disaster response

· Predicting green investment returns in volatile markets

· Designing and monitoring climate policies

Whilst AI certainly holds transformative potential to accelerate the climate transition Stern et al., (2025) caution that “realizing this potential requires targeted public investment, data sharing, and equitable access to AI so no country in the net-zero transition is left behind”.

5. AI and Battery Recycling: Closing the Loop

On the circular economy front, AI has been shown to improve material circularity in the context of lithium-ion battery recycling. Zhou (2025) details how machine learning is being used to:

· Automate sorting of battery components

· Identify cathode types to maximize recovery

· Optimize energy and chemical use in recycling workflows

Such advances boost the recovery rates of minerals like lithium, cobalt, nickel, and graphite, elements critical to the production of net-zero technologies like electric vehicles, solar panels, and wind turbines.

But recycling isn’t just about sustainability. As countries race to secure these critical minerals, recycling becomes a strategic imperative. Most of the world’s mineral supply chains are concentrated in just a few countries, and China in particular has used its dominance in processing and refining to exert geopolitical influence. In this context, being able to recover and reuse materials domestically reduces dependency on politically sensitive supply routes and enhances resource sovereignty.

By embedding AI into recycling workflows, the industry can move toward a more sustainable, closed-loop battery ecosystem – supporting both climate goals and geopolitical stability in an increasingly resource-constrained world.

This Week’s Takeaway

AI is neither inherently green nor destructive – it is what we make of it. Whether it accelerates or undermines our climate goals will depend on how we regulate it, how transparently it operates, and how equitable its benefits and burdens are shared. There is certainly cause to pursue the development of AI for its potential to drive a transformative shift toward a net-zero economy – but this must not come at the expense of the planet.

Thanks for reading