UCT senior lecturer and researcher, Grant Oosterwyk.
Grant, a senior lecturer and PhD candidate in Information Systems at UCT, recently co-authored a paper with international collaborators examining how the rapid expansion of AI infrastructure is generating significant sociotechnical tensions that extend far beyond technological innovation.
Drawing on six expert interviews and an extensive media-based analysis, the research identified five major tensions linked to the growth of AI data centres: the energy paradox, water strain, hyperscaler dominance, sovereignty erosion and urban displacement.
“AI is often framed through a techno-optimistic lens that focuses primarily on innovation and economic growth,” he said. “Our findings show that the rapid growth of AI data centres also creates significant environmental, social and political tensions that require much more critical and integrated attention.”
Energy and water pressure
One of the central findings of the study is the “energy paradox”: AI systems require enormous computational power, driving rising electricity consumption in data centres at a time when many countries are already struggling with energy insecurity and climate-related pressures.
The study also points to increasing water strain caused by AI infrastructure. Data centres rely heavily on water cooling systems to prevent servers from overheating, placing additional pressure on already stressed water resources in many regions.
“The expansion of AI infrastructure is not occurring in isolation. It intersects directly with broader concerns around energy access, climate resilience and environmental sustainability,” Grant said.
Corporate concentration
He noted, “The concentration of AI infrastructure within a small number of multinational corporations risks reinforcing global inequalities and limiting the ability of smaller nations and local governments to exercise control over their own digital futures.”
The concept of “sovereignty erosion” emerged as another major tension identified in the research.
“Our findings suggest that AI infrastructure is increasingly becoming a geopolitical issue. Countries are grappling not only with technological adoption, but also with questions of ownership, control and long-term digital sovereignty,” he said.
He added that the impacts of AI infrastructure are unevenly distributed, with certain communities bearing greater environmental and social costs while receiving limited direct benefits from technological advances.
“These tensions demonstrate the need for a more inclusive and socially responsive approach to AI governance and infrastructure planning,” he said.
Addressing the challenges associated with AI-driven data centre demand will require a more balanced and interdisciplinary approach that moves beyond purely technological or market-driven perspectives.
“Balancing AI’s transformative potential with environmental sustainability and social responsibility will be one of the defining policy challenges of the coming years. Governments, industry leaders and researchers will need to work together to ensure that AI infrastructure development is both resilient and equitable,” he said.
Subscribe to ESG Global newsletter