The rapid growth of the artificial intelligence (AI) industry could lead to an energy consumption level comparable to that of a small country, such as the Netherlands, by 2027, according to a recent study. As tech giants rush to incorporate AI-powered services into their offerings, these applications are shown to consume significantly more power than conventional ones, making online activities more energy-intensive. While the environmental impact of AI is a concern, the study suggests that this scenario could be mitigated if AI’s current growth rate slows down.
AI’s Energy Consumption
AI applications require more powerful hardware than traditional computing tasks. The study, conducted by Alex De Vries, a PhD candidate at the VU Amsterdam School of Business and Economics, makes predictions based on several parameters, assuming certain conditions remain unchanged, such as AI’s current growth rate and the continuous operation of AI chips and servers.
De Vries’s analysis is grounded in the estimate that chip designer Nvidia supplies approximately 95% of the AI processing hardware needed by the sector. By projecting the amount of these computers expected to be delivered by 2027, the study estimates that AI could consume between 85 and 134 terrawatt-hours (TWh) of electricity annually. At the higher end of this range, it equates to the electricity consumption of a small country.
Environmental Impact of AI
The energy-intensive nature of AI systems, particularly those powering chatbots like ChatGPT, necessitates warehouses full of specialized computers, called data centers. These data centers require significant power to operate, and, like conventional equipment, they require cooling systems that consume water. While the study doesn’t factor in cooling energy usage due to limited data transparency from tech firms, experts call for more transparency on this matter.
The study suggests that AI consumption could be considered a new extractive industry due to its significant electricity and water requirements, posing environmental concerns.
AI’s Potential to Address Environmental Challenges
Despite its energy consumption, AI offers promise in addressing environmental challenges. For instance, AI tools can help reduce the number of contrails (vapour trails) left by aircraft, known to contribute to global warming. Furthermore, AI’s potential in the research of nuclear fusion and its role in speeding up advancements is noted. Success in nuclear fusion could lead to a nearly limitless, green power supply, which would have far-reaching environmental benefits.
The study underscores the significant energy consumption of AI applications, which, if left unchecked, could be equivalent to a small country’s electricity usage. However, the environmental impact could be reduced with more conservative AI growth. While AI poses challenges, it also holds the potential to address environmental issues, such as reducing contrails and advancing nuclear fusion research. Balancing AI’s benefits and environmental concerns remains a key challenge for the industry.