That potential reduction would vastly outweigh the estimated emissions generated by AI infrastructure itself, which are projected to range between 0.4 and 1.6 billion tonnes per year by that time. Researchers behind the study emphasize that AI should be viewed not merely as a tool for minor improvements, but as a transformative force capable of turning the global economy into a more sustainable, resilient, and inclusive system. The transition to a net-zero world, they argue, is not a burden but a massive opportunity for innovation and shared progress.
The analysis focuses on three high-emission sectors: electricity, meat and dairy, and passenger transport. Together, these account for nearly half of all global emissions. In these areas alone, AI has the capacity to optimize energy systems, accelerate sustainable food innovations, and improve transportation efficiency.
One major benefit of AI lies in its ability to manage complex systems more effectively. For instance, energy grids that rely on renewables often face inconsistencies due to changing weather conditions. AI can forecast demand and supply fluctuations and help balance energy loads in real time, reducing the need for fossil fuel-based backup power. Studies show AI has already helped increase the efficiency and value of wind power by as much as 20%.
AI is also expediting the discovery of new technologies that are crucial for decarbonization. Nearly half of the emission reductions required to reach global climate goals depend on innovations that are still in development. AI can accelerate this process by simulating, testing, and optimizing materials and systems at a scale far beyond human capability. In clean tech, AI has identified millions of promising new materials for energy storage and conversion. In industry, it’s reducing waste through smarter packaging and logistics, with some platforms eliminating millions of tonnes of unnecessary material use.
Another major opportunity is in behavior change. AI can serve as a guide, helping individuals make lower-emission choices in their daily lives—from food to travel. Navigation systems that suggest fuel-efficient routes, smart home thermostats that adjust energy use, and apps that recommend sustainable alternatives are already making an impact. If widely adopted, these tools could lead to massive cumulative savings in energy and emissions.
AI also plays a vital role in predicting climate trends and the potential impacts of environmental policy. By analyzing global data, machine learning models can forecast changes in weather, sea levels, and agricultural productivity with greater accuracy. These predictions are essential for governments planning long-term infrastructure, risk reduction, and adaptation strategies.
In an increasingly volatile world, AI is also improving disaster preparedness. With the rise of floods, wildfires, and extreme weather events, AI-powered early warning systems are already saving lives. Some tools can forecast floods up to five days in advance, giving communities critical time to prepare and respond.
The numbers reveal that by 2035, AI could deliver the following annual emission cuts:
- Up to 1.8 billion tonnes in the power sector through smarter grid management and renewable energy optimization
- Between 0.9 and 3.0 billion tonnes in food systems by improving plant-based protein technologies
- As much as 0.6 billion tonnes in passenger transport via better battery technologies and shared mobility platforms
However, these benefits will not materialize automatically. The researchers warn that a purely market-driven approach to AI development could result in uneven gains, or even environmental harm. Instead, they advocate for what they call an “active state”—a coordinated public effort to ensure AI is used ethically, equitably, and sustainably.
This includes investing in green AI research, implementing regulations to minimize environmental harm from data centers and hardware, and ensuring that underserved communities worldwide also benefit from AI advancements.
Governments must also guide private innovation in ways that align with climate goals. That means offering incentives for carbon-cutting AI solutions, expanding access to clean technologies, and fostering international cooperation to prevent inequities in AI adoption and climate resilience.
Ultimately, the study paints a picture of AI as a central player in the global climate transition—not a threat, but a tool that, when shaped wisely, can help humanity meet its most urgent challenge. With the right investments and policies in place, artificial intelligence could lead the way to a cleaner, safer, and more sustainable future where people and the planet thrive together.
