Debunking the Hype: Experts Challenge the Link Between Nuclear Fusion and AI's Energy Demands

Artificial intelligence requires a lot of energy, and as companies work to enhance its capabilities, its demand for electricity will only grow. This poses a challenge for an industry that promotes itself as a solution to environmental problems, as it contributes to a significant carbon footprint.

However, Sam Altman, the leader of OpenAI, the company behind ChatGPT, believes that there is a straightforward solution to this issue: nuclear fusion.

Altman has invested hundreds of millions in fusion technology. In recent interviews, he mentioned that fusion is considered the holy grail of clean energy and will be able to meet the huge power requirements of next-gen AI.

In a January interview, Altman emphasized the need for a breakthrough in fusion technology, stating that it is essential along with the expansion of other renewable energy sources. When asked by podcaster Lex Fridman in March about solving AI's energy challenges, Altman once again highlighted the importance of fusion technology.

Nuclear fusion, which is the process that fuels the sun and other stars, is expected to take several more decades before it can be effectively utilized and commercialized here on Earth. Some experts believe that Altman's focus on a future energy breakthrough highlights a broader issue within the AI industry. They have not yet addressed how they plan to meet the rapidly increasing energy demands of AI in the short-term.

The inside of the JET tokamak, which conducted major nuclear fusion experiments in the UK.

The inside of the JET tokamak, which conducted major nuclear fusion experiments in the UK.

United Kingdom Atomic Energy Authority

Related article

Scientists say they can use AI to solve a key problem in the quest for near-limitless clean energy

Alex de Vries, a data scientist and researcher at Vrije Universiteit Amsterdam, highlighted a common trend of "wishful thinking" in climate action. He suggested that it is more practical to focus on current actions rather than relying on uncertain future outcomes.

When asked for a response, a representative from OpenAI did not provide specific answers to the questions posed by CNN. Instead, they referred to comments made by Altman earlier in the year and on Fridman's podcast.

The AI industry is drawn to nuclear fusion for its clear benefits. Fusion is the process of combining two or more atoms to create a denser one, resulting in the release of large amounts of energy.

Unlike other energy sources, fusion does not emit carbon pollution into the atmosphere and does not leave behind long-lasting nuclear waste. This presents an exciting prospect of a clean, secure, and plentiful source of energy.

Recreating the extreme conditions found in the center of the sun on Earth poses a significant challenge. According to Aneeqa Khan, a research fellow at the University of Manchester, the necessary technology may not be available until the latter half of the century.

Khan emphasized that fusion energy is not a timely solution to address the climate crisis. She suggested that in the short term, it is important to utilize existing low-carbon technologies like fission and renewables.

Fission is the process widely used to generate nuclear energy today.

A section of JT-60SA, a huge experimental nuclear fusion reactor at Naka Fusion Institute in Naka city of Ibaraki Prefecture, Japan, on January 22, 2024.

A section of JT-60SA, a huge experimental nuclear fusion reactor at Naka Fusion Institute in Naka city of Ibaraki Prefecture, Japan, on January 22, 2024.

Philip Fong/AFP/Getty Images

Finding sufficient renewable energy to meet the increasing demands of AI in the short term is a challenge, especially as the move away from fossil fuels gains momentum. This is particularly crucial as the global shift toward electrification, encompassing everything from vehicles to heating systems, drives up the need for clean energy.

According to a recent analysis by the International Energy Agency, the electricity consumption from data centers, cryptocurrencies, and AI is projected to double within the next two years. In 2022, this sector accounted for approximately 2% of the total global electricity demand, as reported by the IEA.

The demand for AI is expected to grow significantly from 2023 to 2026, with an exponential increase of at least 10 times.

In addition to the energy needed for producing chips and other hardware, AI also requires substantial computing power for training models by feeding them massive datasets. This power is also needed for utilizing the trained models to generate responses to user queries.

With the advancement of technology, companies are eager to incorporate it into apps and online searches, leading to an increase in computing power needs. A recent report by de Vries estimated that an online search using AI could consume at least 10 times more energy than a standard search.

According to de Vries, the trend in AI is towards larger models that are more energy-intensive, as "bigger is better" in the world of AI. This poses a challenge for sustainability, as the focus on bigger models contradicts the principles of sustainability.

The situation in the US is becoming more serious with energy demand on the rise for the first time in about 15 years, according to Michael Khoo, climate disinformation program director at Friends of the Earth and co-author of a report on AI and climate. He mentioned to CNN that "We as a country are running out of energy."

Part of the reason for this increasing demand is the growth in data centers. A Boston Consulting Group analysis predicts that data center electricity consumption will triple by 2030, which is equivalent to the amount needed to power approximately 40 million US homes.

"We have some tough choices to make," Khoo mentioned, referring to deciding who will receive the energy. This could mean prioritizing thousands of households or a data center that supports advanced AI technology. He emphasized that it shouldn't just be the wealthiest individuals who receive the energy first.

High school students work on "Alnstein", a robot powered by ChatGPT, in Pascal school in Nicosia, Cyprus on March 30, 2023.

High school students work on "Alnstein", a robot powered by ChatGPT, in Pascal school in Nicosia, Cyprus on March 30, 2023.

Yiannis Kourtoglou/Reuters

Many AI companies often overlook two important points when it comes to their energy use. The first point is that AI itself can actually help in addressing the climate crisis.

According to a spokesperson from Microsoft, which collaborates with OpenAI, "AI will play a crucial role in advancing sustainability solutions."

AI is currently being utilized for various purposes such as predicting weather, monitoring pollution, mapping deforestation, and tracking melting ice. According to a recent report by Boston Consulting Group, sponsored by Google, AI has the potential to reduce up to 10% of global warming emissions.

In addition, AI could play a significant role in the advancement of nuclear fusion. In February, researchers at Princeton University announced a breakthrough in using AI to predict potential instabilities in nuclear fusion reactions, a crucial step towards making nuclear fusion commercially viable.

AI companies also say they are working hard to increase efficiency. Google says its data centers are 1.5 times more efficient than a typical enterprise data center.

TOPSHOT - Firefighters work on the zone of a forest fire in the hills in Quilpue comune, Valparaiso region, Chile on February 3, 2024. The region of Valparaoso and Viña del Mar, in central Chile, woke up on Saturday with a partial curfew to allow the movement of evacuees and the transfer of emergency equipment in the midst of a series of unprecedented fires, authorities reported. (Photo by Javier TORRES / AFP) (Photo by JAVIER TORRES/AFP via Getty Images)

Firefighters are seen working on the zone of a forest fire in the hills in Quilpue comune, Valparaiso region, Chile on February 3, 2024. The region of Valparaiso and Viña del Mar, in central Chile, had a partial curfew on Saturday to help evacuees and emergency equipment move around during a series of unprecedented fires, authorities reported. (Photo by Javier TORRES / AFP) (Photo by JAVIER TORRES/AFP via Getty Images)

Related article

The world just marked a year above a critical climate limit scientists have warned about

Microsoft is focusing on researching to determine the energy consumption and carbon footprint of AI. They are also working on enhancing the efficiency of large systems, both in training and application.

According to de Vries, there has been a significant improvement in the efficiency of AI. However, he warns that this does not automatically translate to a decrease in AI's electricity usage.

In fact, history shows that technology and automation may not always lead to reduced energy consumption, according to de Vries. Take cryptocurrency, for example. Despite efficiency gains, energy consumption in cryptocurrency mining has not decreased. When goods and services become more efficient, it often results in increased demand.

In the US, there is a growing political push to examine the environmental effects of AI more closely. Senator Ed Markey recently proposed a bill that would require AI companies to be more transparent about their environmental impacts, specifically focusing on the rising electricity demand of data centers.

Markey emphasized the importance of prioritizing the health of our planet in the development of AI tools. However, there is skepticism about the bill receiving the necessary bipartisan support to become law.

Khoo pointed out that the trend towards more complex and energy-intensive AI is seen as unavoidable, with companies competing to create cutting-edge technology. This has led to the creation of larger models and increased electricity consumption.

"Whenever someone claims to be addressing the issue of climate change, we need to question how they are actually taking action," Khoo explained. "Are they working towards reducing energy consumption each day? Or is it just a distraction tactic?"

Editor's P/S:

The article presents a compelling exploration of the energy challenges posed by artificial intelligence (AI) and the potential of nuclear fusion as a solution. It highlights the increasing demand for electricity as AI capabilities expand, raising concerns about the industry's environmental impact. However, the article also acknowledges the challenges in developing and commercializing nuclear fusion technology, which may not be a timely solution to address the current energy needs of AI. Balancing the potential benefits of AI with its energy consumption requires careful consideration and a focus on efficiency improvements.

The discussion also emphasizes the importance of AI's role in addressing climate change, such as through its use in predicting weather patterns and monitoring pollution. This underscores the need for a comprehensive approach that considers both the environmental impact of AI and its potential to contribute to sustainability. The article raises important questions about the responsibility of AI companies in mitigating their energy consumption and the need for transparency and regulation to ensure that AI development aligns with environmental goals.