The use of artificial intelligence (AI), from navigation on smartphones to industrial applications in robotics and medicine, requires significant amounts of energy. Quantifying the exact energy consumption that AI will require in the future is still a topic of research and poses significant challenges and opportunities for the energy sector.
Morgan Stanley indicates that AI energy demands will skyrocket 70% annually. By 2027, AI could use as much energy as Spain needed to feed itself in 2022. That year, primary energy consumption in Spain was 118,232 ktoe (kilotonnes of oil equivalent), which represents the total amount of energy consumption in the country, measured in a unit that is equivalent to one thousand tons of oil. This growth is a great challenge for data centers.
Data centers are expanding rapidly to handle the growing need to process large volumes of data in areas such as speech, text, images and near-human-like videos. Microsoft, for example, is adding a new data center somewhere in the world every three days.
This expansion of data centers has direct consequences on energy consumption at a national and international level. In the United States, data centers are estimated to use more than 480 TWh of electricity annually by 2035, a tenth of the country’s total consumption and more than double what they will consume this year, according to S&P Global Commodity Insights. One TWh is equivalent to the electricity consumption of approximately 86,000 homes for one year.
Globally, the International Energy Agency (IEA) estimates that global electricity demand for data centers could reach more than 1,000 TWh by 2026, double what was used in 2022 and equivalent to Germany’s total energy consumption.
Who will produce the energy needed for AI?
This increase in energy demand from AI requires significant expansion and modernization of data centers, forcing existing energy infrastructures to adapt and scale. In the United States, limitations on transmission lines and slowness in project planning and approval are key barriers.
The transition to renewable energy, supported by $1 trillion in public incentives to renew the US power grid, is part of the solution. Leading companies such as Google and Microsoft are also committed exclusively to renewable energy in their global operations. However, current renewable energy capacity may not be sufficient to meet growing AI demand, requiring utilities to develop significant new generation and transmission capacity.
Energy demand could benefit several sectors
According to Morgan Stanley, utilities, tasked with expanding and managing energy capacity, would need to scale their operations to meet growing data center demand, which would require significant capital investments. Investments of 5,000 million to 10,000 million dollars and the increase in public budgets between 10% and 20% annually.
Fuel cell manufacturers would also be favored, since their products provide a constant energy source that complements less predictable renewable energies, facilitating their integration into the grid. In addition, nuclear energy companies, which offer clean and uninterrupted energy, would play a crucial role in maintaining the reliability and stability of the grid in a context of growing energy demand.
The natural gas sector, vital for energy generation and as support during the energy transition, would also experience an increase in demand. On the other hand, Power Purchase Agreement (PPA) Providers, who manage contracts with energy sources such as wind and sun, would have the opportunity to secure long-term profitable agreements, benefiting from the guaranteed demand of large electricity consumers. energy.
Finally, interconnection companies, in charge of facilitating the connection of new energy sources to the existing grid, would experience an increase in demand for their services and products, further consolidating the energy infrastructure necessary to support this growth.
