Artificial intelligence, a driver of major innovations, could become a burden for the planet if its energy consumption is not controlled. At Yiaho, we are obviously concerned about the pollution related to AI and its carbon footprint, a topic we have been reflecting on for years. It is therefore essential to share the conclusions and analyses of the Shift Project.
In a report published on October 1, 2025, the Shift Project, a French think tank dedicated to decarbonization and led by Jean-Marc Jancovici, sounds the alarm: without rigorous planning, the electrical footprint of global data centers could triple by 2030, threatening international climate goals.
AI: Exploding Electricity Consumption
The report highlights an alarming projection: the electricity consumption of data centers could increase from 530 terawatt-hours (TWh) in 2023 to between 1,250 and 1,500 TWh in 2030.
This dizzying rise would be largely driven by the growth of AI and the massive use of models like our free online ChatGPT, whose computing needs, particularly for training deep learning models, are particularly energy-intensive.
While in 2023, AI accounted for about 15% of data center consumption, its share could climb to 35, or even 55%, by the end of the decade.
This trajectory, described as “unsustainable” by the Shift Project, jeopardizes commitments made under climate agreements, particularly those of the Paris Agreement. Data centers, already responsible for a significant portion of global greenhouse gas emissions, could become a major obstacle to ecological transition if nothing is done.
Also read on this topic: How to use AI without (too much) polluting the planet? 5 tips for responsible use
Imposing an Energy Cap: The Solution?
Faced with this challenge, the think tank proposes a radical measure: establishing an “energy consumption cap” for the digital sector. This “cap trajectory” would aim to limit the carbon impact of technologies, particularly AI, while ensuring compatibility with carbon neutrality objectives.
According to Shift Project experts, such regulation would require international coordination and ambitious public policies to frame the development of digital infrastructures.
The report also emphasizes the importance of optimizing existing technologies and favoring less energy-intensive solutions, such as more efficient algorithms or data centers powered by renewable energy.
However, these measures will not be enough without global governance to curb the uncontrolled expansion of computing capacities.
A Call for Immediate Action
The Shift Project insists on the urgency of action. It calls on governments, technology companies, and international institutions to collaborate to integrate energy sobriety into AI development.
While tech giants continue to invest massively in infrastructure to support AI, this report highlights a paradox: the technological revolution, often presented as a solution to the challenges of the 21st century, could worsen the climate crisis if it is not regulated.
The Shift Project thus poses a crucial question: will humanity be able to reconcile innovation and environmental responsibility?
Source: TheShiftProject
