Artificial Intelligence Consumes Increasing Amounts of Electricity and Accelerates Investments in Data Centers

The growth of artificial intelligence is increasingly being measured in energy terms. Behind the advancement of generative models, chatbots, and massive data processing platforms lies a high-consumption physical infrastructure: data centers.

According to the International Energy Agency (IEA), the global electricity demand of these facilities grew by 17% in 2025. In the case of centers specifically focused on artificial intelligence, consumption increased by 50% during the same period.

This leap redefines investment decisions, energy planning, and the location of technological infrastructure. In this map, Argentinian Patagonia has begun to gain traction as a possible destination for new large-scale data centers, due to its availability of renewable energy, low population density, cold climate, and connectivity.

AI Accelerates Global Electricity Consumption

According to projections by Gartner for 2026, the total of data centers will consume 565 TWh of electricity this year. Within that total, servers optimized for artificial intelligence will account for 31%, with a year-on-year growth of 84%.

The International Energy Agency estimates that the global consumption of data centers could reach 945 TWh by 2030, almost double that of 2024. This figure shows the pace of expansion of an industry that no longer relies solely on chips, software, or computing capacity, but also on firm and available energy.

The main driver of this growth is hardware. Next-generation graphics processors, known as GPUs, account for a significant portion of the electrical consumption needed to train and run artificial intelligence models.

As a scale reference, the United States Congress calculated that training a large model with eight advanced GPUs for eight hours, with an average utilization of 93%, requires 62 kWh. Multiplied by millions of daily operations, this consumption makes energy one of the main bottlenecks for the sector.

The United States Seeks to Sustain Electrical Infrastructure for AI

The United States concentrates the largest global infrastructure related to artificial intelligence and is already facing tensions due to the impact of this demand on the electrical grid.

President Donald Trump announced a non-binding commitment between energy producers and technology companies to finance or build the electrical infrastructure necessary to sustain the growth of AI. The official goal is to prevent the cost from falling on residential users.

In exchange, the government allows these companies to build their own power plants. However, the announcement did not clear up doubts. Consumer advocacy groups considered the commitment to be symbolic and lacking compliance mechanisms.

Meanwhile, the Energy Information Administration projects that residential electricity rates will increase by 5.1% in 2026 and 2.4% in 2027. This scenario fuels the debate over who should finance the expansion of the infrastructure needed to supply data centers.

Questions are also growing in communities near new projects. Residents from various areas participate in public hearings where they reject installations due to noise, environmental impact, water usage, and fears of an increase in electricity bills.

The sector, represented by the Data Center Coalition, defends the role of these infrastructures by pointing out that they support services such as telehealth, digital education, and banking systems.

While the United States discusses how to sustain the expansion of its network, Argentine Patagonia emerges as one of the regions with the greatest potential to attract next-generation data centers.

In October 2025, OpenAI and the Argentine developer Sur Energy signed a letter of intent to build the largest data center in Latin America in the region. The Stargate Argentina project contemplates an estimated investment of up to $25 billion and a capacity of 500 megawatts.

The scale of the project would be eight times larger than the current largest facility in the region, located in São Paulo. The initiative is evaluating land in the Confluencia department in Neuquén and is supported by the Regime of Incentives for Large Investments (RIGI).

The choice of Patagonia responds to several factors. The region has abundant renewable energy, especially wind farms with capacity factors exceeding 40%, in addition to industrial-scale hydroelectric infrastructure, such as Piedra del Águila, with 1.44 GW installed.

Additionally, there is low population density, available space, and a naturally cold climate, which can reduce cooling costs. In a conventional data center, cooling accounts for between 30% and 40% of total electricity consumption.

Connectivity also appears as a key point. The region has fiber optic and submarine cable connections, a central requirement for facilities that must process and transfer large volumes of information.

OpenAI, Sur Energy, and Tesla are looking at Neuquén.

The Stargate Argentina project was presented as a long-term initiative. Eight months after the announcement, public progress is limited, and OpenAI acknowledges that it is a complex development.

Sam Altman, CEO of OpenAI, stated that the goal is to bring artificial intelligence closer to people throughout Argentina. However, the project still needs to advance in technical, regulatory, and infrastructure definitions.

In parallel, Tesla is evaluating its own data center in Neuquén alongside YPF Luz. This move again highlights the province's appeal as an energy and technology hub, in a context where the availability of electricity has become a decisive factor for the global AI industry.

Neuquén also concentrates the development of Vaca Muerta, adding an additional energy dimension. The province combines hydrocarbon resources, electrical infrastructure, and renewable projects, although it also faces challenges related to water use, networks, and land planning.

Environmental and Regulatory Risks

The interest in new data centers coexists with concrete tensions. Large-scale facilities may require millions of liters of freshwater for cooling, a sensitive issue for Patagonia, where there are areas with water stress and rivers with historically low flows.

This point becomes more relevant due to the coexistence with other resource-intensive activities, such as the development of Vaca Muerta. The discussion about water usage, pressure on the electrical grid, and territorial impact emerges as one of the pending issues for any advancement of projects of this scale.

In terms of employment, the construction of a data center can generate thousands of temporary jobs. However, the permanent operation of a large-scale facility employs between 50 and 100 people, according to the cited data.

Another pending issue is regulation. Brazil requires 100% sustainable energy to grant tax benefits to data centers, and Chile has a National Data Centers Plan 2024-2030. Argentina, on the other hand, still lacks specific regulations for this type of facility.

The tax benefits of the RIGI do not require specific conditions for sustainability, water efficiency, or commitments to local technological development for data centers. This regulatory absence opens a discussion on how to manage energy and resource-intensive investments.

The energy, the new bottleneck of artificial intelligence

The advancement of artificial intelligence modifies the global energy map. The electricity demand from data centers is growing at an accelerated pace and forces the definition of new investments in generation, transportation, cooling, and connectivity.

The United States is trying to sustain the boom without passing the cost onto households. Argentine Patagonia, meanwhile, appears as an attractive region due to its clean energy, climate, and territorial availability.

The challenge for Argentina will be to convert that interest into concrete investments, with rules that allow for the orderly use of energy, water, and infrastructure. The opportunity is linked to attracting large-scale projects. The condition will be to define who pays and who benefits from the energy that artificial intelligence needs to continue growing.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.