By Mandy Hattingh, Director: Energy & Environmental, NSDV
Artificial Intelligence (AI) and the pursuit of Artificial General Intelligence (AGI) are rapidly changing the global landscape, and South Africa cannot afford to be left behind. While AI has existed in various forms for decades, it is only recently, with advances in machine learning and the discovery of large language models, that AI has come to the forefront of societal, ethical, and governance discussions. These discussions, particularly in the context of Environmental, Social, and Governance (ESG) considerations, are especially critical for South Africa, where energy inequality, water scarcity, and social inequality are prevalent.
Environmental Impacts
The environmental implications of AI development, particularly in the South African context, are significant. The most commonly noted environmental impacts of AI and the data centres supporting its development are the substantial volumes of electricity and water they consume.
AI systems require immense data processing capabilities, driving the expansion of data centres, which, in turn, consume large amounts of electricity. In a country where energy reliability remains a pressing issue, this additional demand may exacerbate the ongoing energy crisis and increase greenhouse gas emissions, as South Africa still heavily relies on its ageing coal-fired power stations. In April 2024, the World Economic Forum estimated that the computational power required to sustain the rise in AI globally doubles every 100 days. Subsequent developments in large language models and generative AI have only accelerated this trajectory, with both training and real-time inference workloads driving exponential growth in global compute and electricity demand. This highlights the urgency of addressing energy capacity constraints, particularly where those demands are met by carbon-intensive fossil fuels.
Renewable energy presents a promising solution to power AI, but intermittency issues mean that it must be hybridised with either another renewable energy source, or with storage solutions to ensure an uninterrupted power supply. However, as AI usage expands, industry is concerned that renewables alone may not suffice to meet energy demands. Consequently, companies such as Google, Amazon, and Microsoft have invested in nuclear energy, recognising it as a reliable, carbon-free source capable of scaling with AI’s needs. This highlights the importance of South Africa exploring a mix of sustainable energy sources while also potentially considering advanced solutions, such as nuclear energy, to meet its AI-related energy demands.
In addition to energy, data centres consume vast amounts of water, primarily for cooling. South Africa, being a water-scarce country, faces a substantial challenge in balancing technological advancement with responsible water usage. Accurately tracking water consumption in data centres remains challenging because many operators do not systematically measure or disclose their water use. According to 2025 industry analyses, fewer than half of data centre operators actively monitor water usage metrics, and reporting standards vary widely across firms and jurisdictions. For context, a single large data centre can use up to 2 million litres of water per day. This is comparable with the daily water needs of about 6 500 households.
In South Africa, where recent droughts and water restrictions have heightened concerns, ensuring AI development does not worsen the water crisis is crucial. To mitigate this, industry leaders are increasingly adopting cooling technologies that minimise water use. However, more widespread tracking of water usage in the industry is necessary to assess and reduce the negative environmental footprint of data centres effectively.
Moreover, data centres generate significant volumes of electronic waste (e-waste) due to frequent hardware upgrades. In 2022, the world generated a record 62 million tonnes of e-waste — equivalent to about 7.8 kg per person globally, yet only 22 per cent of it was formally collected and recycled. E-waste volumes are projected to rise to roughly 82 million tonnes by 2030. Although e-waste is regulated in South Africa under the Extended Producer Responsibility Programme, informal and unsafe recycling practices remain prevalent. South Africa’s reliance on informal recycling, coupled with limited monitoring and enforcement, poses a risk that AI-driven e-waste could exacerbate environmental and health issues unless stronger measures are implemented. To avoid compounding the e-waste problem, South Africa must aim to improve its recycling capacity, educate stakeholders, and enforce monitoring and regulations under the applicable legislation.
While AI and AGI present resource challenges, they also hold significant potential for environmental benefits. AI can enhance sustainability, resource efficiency, and climate change mitigation by optimising key industries such as energy, agriculture, and transportation. For example, AI-driven smart grids can manage energy consumption more efficiently, reducing waste and improving the integration of renewable energy sources. This helps reduce reliance on fossil fuels and lowers carbon emissions. AI also plays a vital role in predictive climate modelling, providing policymakers with tools to make informed decisions on mitigation and adaptation strategies. By analysing large datasets, AI can identify climate trends, predict natural disasters, and improve responses. Various industry analyses suggest that AI could enable emissions reductions of between 5 and 10% of global greenhouse gas emissions by 2030 through optimisation of energy systems, transport networks, agriculture, and industrial processes. However, the net climate impact of AI will depend heavily on how its underlying energy demand is met.
AGI takes these benefits further by potentially designing new sustainable technologies, such as advanced carbon capture methods and circular economy systems that minimise waste. AGI could accelerate innovations like nuclear fusion or more efficient battery storage, speeding up the transition away from carbon-intensive fuels and supporting long-term environmental sustainability.
Social Impacts
South Africa remains one of the most unequal societies globally, and AI has the potential to exacerbate this divide. On one hand, AI is expected to revolutionise industries and create competitive advantages for countries, regions, and individuals that have the resources and education to participate in the AI economy. On the other hand, those who lack access to these tools could be left behind.
Read the full story in the 8th edition of ESG: The Future of Sustainability 8th edition
Mandy Hattingh is the Director: Energy & Environmental, NSDV
