ChatGPT CEO proposes nuclear fusion as solution to AI energy needs, but experts disagree

Artificial intelligence requires significant energy, and as companies strive to advance it, its electricity consumption will rise. This poses a challenge for an industry promoting itself as a tool for environmental conservation due to its large carbon footprint.

Sam Altman, CEO of OpenAI, the creator of ChatGPT, believes there's a solution: nuclear fusion. Altman has personally invested substantial amounts in fusion and has indicated in recent interviews that this futuristic technology, considered the pinnacle of clean energy, will eventually supply the vast power needs of future AI.

Altman emphasized the necessity of a breakthrough, stating, "There's no way to get there without fusion," while also advocating for the expansion of other renewable energy sources in a January interview. In a March conversation with podcaster and computer scientist Lex Fridman about solving AI's "energy puzzle," Altman once again highlighted the importance of fusion.

However, nuclear fusion, the process that powers the sun and stars, is likely many decades away from being fully developed and commercialized on Earth. Some experts view Altman's focus on a future energy breakthrough as indicative of the broader failure of the AI industry to address the immediate challenge of meeting AI's escalating energy requirements.

According to Alex de Vries, a data scientist and researcher at Vrije Universiteit Amsterdam, this perspective reflects a broader inclination toward "wishful thinking" regarding climate action. "It would be more practical to concentrate on current possibilities and actions rather than relying on potential future developments," he told CNN.

An OpenAI spokesperson declined to address specific questions from CNN, instead referring to Altman's remarks from January and on Fridman's podcast.

The AI industry is attracted to nuclear fusion for its potential. Fusion involves merging two or more atoms to create a denser one, releasing vast amounts of energy in the process.

Unlike traditional energy sources, fusion doesn't emit carbon dioxide and produces no long-lasting nuclear waste, presenting an appealing vision of a clean, safe, and abundant energy solution.

However, "replicating the conditions at the core of the sun on Earth is an immense challenge," and the technology is unlikely to be viable until the latter part of this century, according to Aneeqa Khan, a nuclear fusion research fellow at the University of Manchester.

Khan stated, "Fusion is already too late to address the climate crisis," emphasizing the need to prioritize existing low-carbon technologies like fission and renewables in the short term. Fission is the process currently used for nuclear energy production.

The challenge lies in securing adequate renewable energy to meet the increasing energy demands of AI in the short term, without resorting to fossil fuels, which contribute to global warming. This is particularly challenging as the global shift toward electrifying vehicles and heating systems drives up the demand for clean energy.

According to a recent analysis by the International Energy Agency (IEA), electricity consumption from data centers, cryptocurrencies, and AI could double over the next two years. In 2022, this sector accounted for approximately 2% of global electricity demand, as per the IEA.

The analysis forecasts exponential growth in AI-related demand, with at least a tenfold increase expected between 2023 and 2026. In addition to the energy needed to manufacture chips and other hardware, AI requires significant computing power for both training models—by processing vast datasets—and responding to user queries.

As AI technology advances, companies are hurriedly incorporating it into applications and online searches, leading to a surge in computing power requirements. According to a recent report by de Vries on AI's energy impact, an AI-powered online search could consume at least ten times more energy than a standard search.

De Vries noted that there is a prevailing belief in the AI industry that "bigger is better," which drives companies to adopt large, energy-intensive models. However, he emphasized that this notion conflicts with sustainability principles, stating that "bigger is better is just fundamentally incompatible with sustainability."

The energy situation is particularly critical in the United States, where energy demand is increasing 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. "We as a country are running out of energy," Khoo told CNN.

The surge in data centers is a major driver of this demand. A Boston Consulting Group analysis projects that data center electricity consumption will triple by 2030, equivalent to the energy needed to power approximately 40 million US homes.

Khoo emphasized the need to make tough decisions about energy allocation, whether to prioritize powering thousands of homes or a data center supporting next-generation AI. He stressed that energy access should not be limited to the wealthiest individuals.

Many AI companies believe that concerns about their energy consumption overlook two key points: First, AI itself can be a powerful tool in addressing the climate crisis.

According to a Microsoft spokesperson, "AI will be a powerful tool for advancing sustainability solutions." Microsoft has a partnership with OpenAI.

AI technology is already being used to predict weather patterns, monitor pollution levels, map deforestation, and track melting ice. A recent report by the Boston Consulting Group, commissioned by Google, suggested that AI could help reduce up to 10% of the carbon emissions that contribute to global warming.

Additionally, AI could play a role in advancing nuclear fusion. In February, scientists at Princeton announced a method for using AI to forecast potential instabilities in nuclear fusion reactions, which represents progress in the journey toward commercializing fusion energy.

AI companies are also focusing on improving efficiency. Google, for example, states that its data centers are 1.5 times more efficient than the average enterprise data center.

A Microsoft spokesperson stated that the company is "investing in research to measure the energy use and carbon impact of AI, while working on ways to enhance the efficiency of large systems in both training and application."

There has been a significant increase in AI's efficiency, according to de Vries. However, he cautioned that this does not necessarily mean a decrease in AI's electricity demand.

In fact, historical trends in technology and automation suggest that the opposite may be true, de Vries noted, citing cryptocurrency as an example. He explained, "Efficiency gains have never reduced the energy consumption of cryptocurrency mining. When we make certain goods and services more efficient, we often see increases in demand."

In the United States, there is some political momentum to scrutinize the environmental impacts of AI more closely. In February, Senator Ed Markey introduced legislation aimed at requiring AI companies to be more transparent about their environmental impacts, including the rising electricity demand of data centers.

Markey stated, "The development of the next generation of AI tools cannot come at the expense of the health of our planet." However, few expect the bill to receive the bipartisan support necessary to become law.

Meanwhile, the development of increasingly complex and energy-intensive AI is viewed as inevitable, according to Khoo, with companies engaged in an "arms race" to create the next breakthrough. This trend leads to the development of larger models and higher electricity consumption.

Khoo emphasized the importance of questioning claims about solving the problem of climate change. He stated, "So I would say anytime someone says they're solving the problem of climate change, we have to ask exactly how are you doing that today? Are you making every next day less energy-intensive? Or are you using that as a smokescreen?"