Issues

Sustainability, Challenges, and Opportunities

Artificial Intelligence (AI) is reshaping industries, transforming how information is processed and utilized. Its ability to generate complex written, visual, and musical outputs quickly is revolutionizing business practices, healthcare, and innovation. However, while its potential is immense, AI adoption carries significant sustainability challenges.

Energy Demands and Growing Power Consumption

AI’s increasing reliance on powerful data centres comes with an extraordinary energy cost. The World Economic Forum reports that the computational power required to sustain AI doubles every 100 days. Currently, data centres processing AI consume 1–2% of global electricity, a figure predicted to rise to 3–4% by 2030.

Moreover, a ChatGPT query requires approximately 2.9 watt-hours (Wh) of energy compared to 0.3 Wh for a typical Google search. The number of global users engaging with AI tools was over 250 million in 2023 and is projected to reach 700 million by 2030. ChatGPT alone set a record in October 2024 with 3.7 billion visits worldwide—a growth of over 115% year-on-year.

Data centres not only use vast amounts of electricity but also place increasing demands on water supplies for cooling. For example, an average of 25 million litres of water is required annually for every megawatt (MW) of processing capacity. Amsterdam’s 950 MW of installed data centre capacity exemplifies the escalating environmental footprint, with AI-related infrastructure potentially consuming as much water as Denmark within a decade.

In January 2025 both the UK and USA announced new initiatives to dramatically increase their capacity to support AI infrastructure.

Environmental Impact and E-Waste Challenges

The environmental impact of AI extends beyond energy and water use. The industry depends on raw materials and generates substantial electronic waste (e-waste) containing toxic substances like mercury and lead. Without proper mitigation, this adds another layer of environmental strain, amplifying AI’s cost to the planet.

Advancements in Sustainability Efforts

Despite its challenges, the AI sector is taking steps toward sustainability. Several tech giants have committed to using renewable energy for their data centres:

   •       Apple: Operates data centres on 100% renewable energy, saving 12 million gallons of water annually since 2023.

   •       Microsoft: Aims to achieve 100% renewable energy usage by 2025 and become water-positive by 2030.

Within the European Union, data centres currently consume around 3% of energy. However, regulations now require operators to track energy efficiency, cooling effectiveness, and renewable energy use. In the UK, where data centres are classified as Critical National Infrastructure, measures are in place to reduce their impact on electricity and water supplies.

Deep Seek: In January 2025 a new ‘free’ AI programme developed by China was announced. It is claimed that it is significantly smaller, cheaper and better than the current market leader ChatGPT. It does not need the huge LLM infrastructure of current large AI programmes. This has sent shock waves across the AI international community. Apple is also launching Apple Intelligence which is also smaller local AI. At the time of writing (January 2025) it is not possible to predict how the momentum towards smaller AI will have an impact on world resources.

In January 2025 the UK government announced a new plan to use AI to drive an action plan for change and to help turbocharge growth and boost living standards. A few weeks later the USA made a similar commitment to develop the infrastructure required for the evolution of AI. This suggests that the cautionary and defensive responses of governments last year are giving way to a new AI race.

Making Sustainable Choices about AI

Decisions regarding the use of AI must account for its energy, environmental, and material costs. By staying informed and prioritizing sustainable practices, educators, businesses, and governments can leverage AI responsibly while minimizing its impact on the planet.