Fynsa
AI energy challenges
Artificial Intelligence, with its ability to process data at unprecedented speed and scale, is transforming many aspects of our lives and businesses. However, this revolution is not without costs. Every action performed by AI systems, from web searches to complex data processing, requires a considerable amount of energy. According to the International Energy Agency (IEA), a simple search in an AI model such as ChatGPT can consume up to ten times more electricity than a conventional Google search.
Generative AI is trained with a significant amount of data through graphics processing units (GPUs) and relies on data centers with high-capacity servers. The power demand of data centers data centers worldwide, according to Morgan Stanley estimates, will increase by 318 TWh between 2024 and 2027. By way of comparison, that increase is equivalent to the total annual energy consumption of Indonesia, or the energy generated by 39 large nuclear reactors.
This growing energy demand is exacerbated by alarming projections from industry leaders, such as those of ARM CEO Rene Haas. He warns that, without significant efficiency improvements, the data centers dedicated to AI could consume as much as 25% of the total U.S. energy supply by 2030. This compares with a modest 4% today.
The role of nuclear energy as a sustainable solution
Against this backdrop, a crucial question arises: How can we meet this growing demand for energy in a sustainable and reliable manner? The answer, according to many experts, lies in nuclear energy, as it stands out as one of the cleanest and most efficient energy sources available.
This is one of the reasons why uranium prices, the main input for nuclear power generation, have reached their highest levels since 2007 this year.. This reflects market expectations that nuclear-related companies will benefit from this AI boom.
Unlike its renewable counterparts, such as solar or wind power, nuclear power is not subject to the constraints of sunlight or wind availability, nuclear power is not subject to the limitations of sunlight or wind availability. It is a highly reliable baseload source, capable of providing constant energy production without significant greenhouse gas emissions. In addition, it has a high energy density, which means that it can pIt can generate large amounts of energy in a small amount of space, which iswhich is crucial in a world where land is scarce and demand is high.
Technological advances and challenges
While it is true that nuclear power presents challenges of its own, such as nuclear waste management and safety concerns, continued technological advances are effectively addressing these concerns. New generation nuclear reactors promise improved safety, efficiency and waste management, making nuclear power an even more attractive option in today's energy landscape.
A brighter energy future
In summary, the convergence of Artificial Intelligence and growing energy demand poses significant challenges to our energy infrastructure. However, nuclear power emerges as a promising and almost indispensable solution to meet this growing demand in a sustainable and efficient manner.
By María Paz Benavente, CFA, MiF LBS | Partner and Co-Head of Research at DVA Capital
