Will AI Cause the Next Major Energy Crisis?

Generative AI is driving a rapidly increasing energy consumption in the world's data centers, challenging tech companies' climate goals. Fossil-free electricity is becoming scarce, and energy-efficient AI could be the solution.

Share
Will AI Cause the Next Major Energy Crisis?

The world's data centers are consuming increasing amounts of electricity, and according to the International Energy Agency (IEA), data centers globally will consume twice as much electricity in 2026 as in 2022. This increased consumption is driven mainly by the need to process AI, especially generative AI, which creates content such as text, images, video, and music.

Generative AI and Its Energy Demands

Generative AI, used in chatbots and other AI services, requires enormous amounts of energy. When users ask a chatbot for a recipe or search for information, it requires far more data processing than a regular search tool, such as when you do a Google search. This energy-intensive process begins already at the training stage of the AI models, where the massive amounts of data needed for AI to "think" require extensive data processing. Today, these are mostly run on graphics cards that require high energy input. If you've ever wondered why your electricity bill goes up at home when the kids are playing on good gaming PCs... multiply that by the enormous number of graphics cards used for AI, and you'll understand the amount of energy required.

Societal Risks

AI technology's energy-guzzling nature risks raising energy costs for everyone, as demand for electricity increases and prices rise. Households and businesses may need to pay more for electricity, which could lead to a higher cost of living and affect economic stability both globally and in Sweden.

The rapidly increasing electricity consumption from data centers could strain Sweden's power grid and create bottlenecks, especially during periods of high demand. This could lead to more power outages and reduced reliability in the electricity supply, affecting both individuals and businesses. Even though Sweden is investing in fossil-free electricity, the increased energy consumption from AI data centers puts greater pressure on quickly scaling up green energy production. If production can't keep pace, Sweden could be forced to use more fossil energy, which would make it harder to meet its climate goals.

Climate Goals and Energy-Efficient AI

Several major tech companies have set net-zero emissions targets for 2030, which is becoming a challenge in light of AI development's energy demands. This has led to an increased focus on supplying data centers with fossil-free electricity. Google recently invested $1.1 billion in an AI data center in Finland, drawn by the country's access to green electricity, primarily wind power. Microsoft, for its part, has invested $3.2 billion in expanding its data centers in Sweden, emphasizing the country's suitability from an energy perspective.

What Are the Challenges, and Possible Solutions?

Despite these investments, fossil-free electricity is becoming scarce, as production can't keep pace with growing demand. To keep AI from contributing to bottlenecks in the energy supply, a number of measures are needed:

- Develop energy-efficient AI models
- Develop AI-specific hardware
- Education and awareness
- Improve data center efficiency
- Energy management and storage
- Distributed computing
- Incentives and legislation