AI industry’s energy consumption could rival that of the Netherlands by 2027, warns a study


  • AI industry’s energy consumption could rival that of a small country by 2027.
  • Big tech companies are rapidly adopting AI-powered services, increasing energy demand.
  • Lack of data transparency hinders precise predictions regarding AI’s environmental impact.
  • AI necessitates robust hardware and specialized chips, primarily supplied by Nvidia.
  • Urges judicious use of AI technology to mitigate energy consumption.
  • AI systems in data centers demand significant power and water for cooling.
  • Growing demand for AI infrastructure poses challenges for data center operators.
  • AI has the potential to address environmental challenges, such as reducing contrail formation and advancing nuclear fusion research.

Main AI News:

The Artificial Intelligence (AI) industry is on the verge of becoming an energy behemoth, potentially rivaling the energy consumption of an entire nation by 2027, according to a recent study. As the technology landscape evolves, big tech giants have eagerly integrated AI-powered services into their offerings since the emergence of ChatGPT. However, these AI applications are voracious power consumers, significantly intensifying the energy demands of online activities.

The environmental impact of AI is a topic of concern, but the study suggests a glimmer of hope if the sector’s rapid expansion can be curtailed. Nevertheless, it’s essential to acknowledge that the speculative nature of such predictions arises from the tech industry’s reluctance to disclose sufficient data for precise forecasts.

One undeniable truth is that AI necessitates more robust hardware than traditional computing tasks. Alex De Vries, a Ph.D. candidate at the VU Amsterdam School of Business and Economics, conducted the study. He based his calculations on certain constants, including the growth rate of AI, the availability of AI chips, and the continuous operation of servers at peak performance. Notably, Nvidia, the chip designer, is estimated to supply approximately 95% of the AI processing equipment required by the industry. By extrapolating the expected number of these computers to be in operation by 2027, De Vries approximated an annual energy consumption range for AI between 85-134 terawatt-hours (TWh) of electricity.

At the upper end of this range, we’re looking at an electricity consumption level comparable to that of a small nation. To put it in perspective, this accounts for roughly half a percent of the total global electricity consumption.

One crucial point Mr. De Vries emphasized is the judicious use of AI technology, emphasizing its deployment only where absolutely necessary. His peer-reviewed study, published in the journal Joule, calls for a pragmatic approach to AI’s energy consumption.

AI systems, such as the massive language models driving chatbots like OpenAI’s ChatGPT and Google’s Bard, rely on specialized computer warehouses known as data centers. These data centers, in addition to being power-hungry, require extensive cooling systems, which often involve significant water usage. Unfortunately, many tech giants do not disclose specific data on energy consumption or water usage related to cooling, highlighting the need for increased transparency in the sector.

The demand for AI-powered computing is skyrocketing, leading to an exponential increase in energy requirements for server cooling. Data center companies are experiencing a surge in inquiries about hosting AI infrastructure, with AI-specific racks consuming substantially more power than standard server racks. In some cases, AI rack power demands are roughly 20 times greater, posing a significant challenge for data center operators.

Despite these energy-intensive challenges, some believe that AI could be part of the solution to environmental issues. Recent experiments by Google and American Airlines demonstrated that AI tools can reduce the formation of contrails from aircraft, which contribute to global warming. Furthermore, government investments in AI for nuclear fusion research could accelerate progress in achieving a green and virtually limitless energy source, potentially revolutionizing the energy landscape.


The soaring energy consumption of the AI industry, poised to rival that of a small nation, underscores the need for responsible and efficient AI deployment. Data transparency remains a challenge, but it’s evident that AI’s impact on energy resources is substantial. Market players should consider sustainable practices and the judicious use of AI technology. While energy-intensive, AI also holds promise in addressing environmental concerns, potentially influencing the market’s trajectory toward more environmentally friendly solutions.