Artificial intelligence (AI) is revolutionizing numerous aspects of society and the economy, from task automation to improving decision-making. Companies and governments around the world are quickly adopting AI technologies to optimize processes, increase efficiency, and create new business opportunities.
In sectors such as healthcare, agriculture, logistics, and education, AI is transforming how we operate and make decisions, offering more precise and personalized solutions.
Furthermore, AI is driving innovation in products and services, creating entirely new markets and reshaping the global competitive landscape. However, this AI revolution brings with it a significant challenge: massive energy consumption.
The exponential rise in energy demand from data centers, driven by the development and deployment of AI, has led the International Energy Agency (IEA) to convene a global summit in Paris on December 5th.
As AI becomes more integrated into our lives, data centers—which are the backbone of this technology—require enormous amounts of energy to operate. This phenomenon has raised concerns about the sustainability and capacity of our energy infrastructure to support this growing demand.
In this article from ITD Consulting, we explore the details of this AI phenomenon, energy consumption projections, and potential solutions proposed by various sectors.
From nuclear energy to renewable sources and advancements in energy efficiency, we discuss strategies that could help mitigate the environmental impact and ensure a stable and sustainable energy supply for the future of AI.
The Growing Energy Demand of AI
The growth of AI has been meteoric, and with it, its demand for energy resources. As a result, it is expected that by 2030, energy consumption in data centers will increase by 200% due to AI.
Some less optimistic estimates suggest that this consumption, driven by AI, could even increase sixfold. This situation has prompted the organization to convene the first Global Conference on Energy and Artificial Intelligence, where strategies will be discussed to address this challenge of AI growth.
Data and Projections
To understand the magnitude of the AI issue, it is useful to compare current and future figures. In 2022, global energy consumption by data centers was approximately 460 terawatt-hours (TWh), representing around 2% of total electricity usage. By 2026, this consumption is predicted to exceed 1,000 TWh, equivalent to the total energy consumption of Japan.
In contrast, cryptocurrencies, known for their high energy consumption, are expected to use about 160 TWh by 2026, a significantly lower figure than that forecasted for AI.
Impact on Energy Supply
The increase in energy demand due to AI has raised concerns about the ability of current energy systems to meet these needs. Elon Musk has warned about a potential "energy drought," indicating that AI could run out of transformers by 2025.
This concern is based on the observation that cloud computing for AI is growing at an unsustainable rate, increasing tenfold every six months.
Nuclear Energy as a Solution
Given the magnitude of the challenge, some experts and tech industry leaders have suggested turning to nuclear energy. Jennifer Granholm, U.S. Secretary of Energy, and Sam Altman, CEO of OpenAI, have advocated for this option.
Nuclear energy has the potential to generate large amounts of consistent power, which is essential to meet the demand from data centers. One example of this push for nuclear energy is Microsoft’s project to generate power from nuclear fusion near the Columbia River in Washington state.
While physicists predict that this goal is difficult to achieve, the company and its partners are focused on this advancement, hoping that nuclear fusion will become a viable solution by 2028.
Experimental Clean Energy Projects
In addition to nuclear fusion, major tech companies are exploring other experimental clean energy projects. These include small nuclear reactors, geothermal energy, and other innovative technologies.
For example, Microsoft is developing small nuclear reactors that could be directly connected to individual data centers. This approach could reduce reliance on fossil fuels and minimize carbon emissions.
Challenges and Criticisms
Despite technological advancements and the promises of large companies, the increase in energy consumption driven by AI has led to a resurgence in the use of fossil fuels. Coal plants, often considered obsolete, are being revitalized to meet the growing demand.
This phenomenon of AI has sparked criticism from environmental organizations and experts, who argue that the rise of AI is delaying the transition to cleaner energy sources.
Tamara Kneese, Director of Projects at Data & Society, has accused the tech industry of using "confusing math" in its claims about environmental impact. According to Kneese, the resurgence of coal plants due to AI should alarm those concerned about the environment.
The Energy Race
The competition among tech companies to develop larger and more powerful data centers has created a true arms race in the energy sector. These centers, which often consume as much energy as a medium-sized city, are forcing energy providers to find new solutions to meet this demand.
In regions like Salt Lake City, the pressure to supply enough energy has led to delays in shutting down coal plants and a reconsideration of investments in clean energy. In Omaha, for example, a coal plant that was set to close in 2022 will continue operating at least until 2026 to support the demand from new data centers for Google and Meta.
Future Solutions and Potential Advances
Despite the current challenges, there is hope that technological advancements and new energy sources can provide long-term solutions. Google, for instance, has started a geothermal project in Nevada that could provide a significant amount of energy in the future.
However, these projects are still in the early stages, and their large-scale impact remains to be seen, given the urgency driven by the exponential growth of AI.
Microreactors and Modular Energy
Another promising solution is microreactors and small modular reactors (SMRs). Companies like NuScale Power are developing SMRs that could generate 77 megawatts of clean energy continuously, enough to operate data centers independently of the electrical grid.
However, these projects also face regulatory and technical challenges before they can be implemented on a large scale.
Sam Altman and Bill Gates, both advocates for nuclear energy, are investing in innovative technologies such as advanced fission reactors and nuclear fusion. These technologies could revolutionize energy production if they manage to overcome the technical and regulatory hurdles.
The growing energy demand of AI presents a significant challenge for global energy systems. As AI expands and integrates into a variety of sectors, the need for energy to power data centers and smart devices is also increasing exponentially.
This AI phenomenon puts considerable pressure on existing energy infrastructures, which are already facing challenges in terms of capacity and sustainability. The International Energy Agency (IEA) and other organizations are working hard to find viable solutions that will meet this demand without compromising the environment.
In this context, clean energy, nuclear energy, and experimental projects emerge as crucial components of the global response to the AI challenge. However, it is essential that these advances in energy supply are made in a sustainable and efficient manner.
Clean energy sources like solar and wind offer significant potential to reduce the carbon footprint of AI, but they require considerable investments in infrastructure and technology to be truly effective. Nuclear energy, on the other hand, can provide a constant and abundant energy source, but it also raises concerns about waste management and safety.
Experimental projects such as nuclear fusion and carbon capture technologies represent promising future innovations, although they are still in the early stages of development and face numerous technical and economic challenges.
The integration of these various energy sources, alongside improvements in energy efficiency for data centers and devices, will be critical to addressing the growing demand in a sustainable way.
The future of AI and energy is intrinsically linked. The ability to develop and maintain efficient and sustainable energy systems will determine the long-term success of artificial intelligence and its impact on society.
The global summit in Paris will be a crucial step in this direction, bringing together experts and leaders from various fields to discuss and plan a path toward a more sustainable energy future. Topics such as international cooperation in research and development, the implementation of favorable energy policies, and the promotion of investments in innovative technologies are expected to be addressed at this summit.
The outcome of these discussions and agreements will be key to ensuring that AI growth is not only technologically advanced but also ecologically responsible and beneficial to the planet. If you'd like to learn more about the impact of AI on energy capacity and how it may affect your business, reach out to us at [email protected]. We have a team dedicated to providing tailored solutions to help you.
EN 
ES