Rebellion against artificial intelligence: protests against data centers grow in the US

As reported The Wall Street Journalpublic rejection against AI is growing with unusual speed and is already expressed in surveys, neighborhood protests, local electoral campaigns and blockades of data center projects. The American newspaper reconstructed cases in which concern about artificial intelligence stopped being concentrated in experts, regulators or workers in the technology sector and began to settle in communities that fear paying its indirect costs.

The episode that opened the WSJ note was a symbolic scene. Eric Schmidt, former CEO of Google, spoke about artificial intelligence during a graduation ceremony at the University of Arizona and received boos from the audience. His message aimed to present AI as a faster and deeper technological transformation than previous ones, but the reaction showed that a part of the audience no longer hears that argument as a promise, but as a threat.

Resistance brings together different concerns. Some users see AI as putting pressure on jobs. Others look at the impact on education, the mental health of children or the circulation of false content. Added to this is a less abstract and increasingly sensitive point: the infrastructure that AI needs to function consumes enormous amounts of energy, requires cooling and multiplies the demand for new data centers.

The phenomenon is no longer limited to a cultural debate about ChatGPT or image generators. The fight enters the field of energy, territory, rates and local politics. And that is why it also resonates outside the United States, in countries that seek to attract artificial intelligence investments, such as the Argentinawhere OpenAI announced last year an ambitious data center project of which few concrete details are still known.

Data centers are the physical infrastructure behind artificial intelligence. Servers, specialized chips, storage systems, high-speed networks and cooling equipment are housed there. In the case of generative AI, this infrastructure becomes even more demanding because the models require enormous processing capacity both to train and to answer millions of queries daily.

This growth explains why the discussion stopped being only technological. The International Energy Agency projects that global electricity consumption in data centers will double by 2030 and reach around 945 TWh, with AI as one of the main drivers of this increase. In the United States, the agency estimates that data centers will explain almost half of the growth in electricity demand between now and 2030.

In this context, neighborhood rejection gains strength. A Gallup poll published in May 2026 showed that 71% of Americans opposes the construction of AI data centers in its area. The data is relevant because it even exceeds local opposition to a nuclear plant, which in the same measurement was 53%.

The WSJ also noted that the unrest has already had electoral and legislative effects. In different communities there were attempts to stop or prohibit new data centers, and in some cases the approval of these projects became a political cost for local officials. The opposition mixes demands for the use of energy and water, fear of rate increases, criticism for opaque agreements and doubts about the real benefits in permanent employment.

More extreme episodes also appear. The WSJ report mentions a federal complaint against a young man from Texas accused of throwing a molotov cocktail against the home of Sam Altman, CEO of OpenAI, and having made threats at the company’s headquarters in San Francisco. He also cites the case of an Indianapolis councilman who had approved a data center and then found shots fired at the door of his house and messages against those projects.

The reaction comes at an uncomfortable time for the industry. OpenAI, Anthropic, Google, Meta, Amazon, Microsoft and other players increasingly need more computing power to sustain their models and services. That race involves multibillion-dollar investments, new energy deals and global competition for chips, land, cheap energy and local permits.

The problem for companies is that the innovation narrative is no longer enough to explain the deployment. When a community sees that a data center can require as much electricity as a medium-sized city, the debate changes. It is no longer just asked what artificial intelligence can dobut rather who pays for the energy, who receives the benefits, what impact is left on the territory and what controls exist before approving a large-scale work.

The American discussion has a local mirror. In October 2025, OpenAI and Sur Energy announced a letter of intent to explore building an artificial intelligence data center in Argentina. The project was presented as Stargate Argentina and, according to the Government, it could reach an investment of up to 25,000 million dollars and a capacity of up to 500 MW.

The announcement was politically strong. The Casa Rosada presented it as an initiative capable of placing the country at the forefront of the global artificial intelligence ecosystem. OpenAI, for its part, published that Argentina had the potential to host the first Stargate project in Latin America, with Sur Energy as an energy and infrastructure developer, and with the possibility of adding a cloud infrastructure developer.

Stargate is the name of OpenAI’s global infrastructure plan. The initiative was announced in January 2025 with SoftBank, Oracle and MGX, with the intention of investing up to $500 billion in AI infrastructure over four years. In this framework, the Argentine chapter was presented as a possible Latin American leg of this expansion.

But since the announcement, public information about the project’s progress has been scarce. In May 2026, Buenos Aires Times published that, more than six months after the presentation, The project had not been formally submitted to the Government under the regime of incentives for large investments. The note cited the Ministry of Economy and specialists who monitor the RIGI, and noted that there was no reliable data on deadlines, precise location or administrative status of Stargate Argentina.

This information gap is important due to the type of work under discussion. A data center of up to 500 MW It is not a common technological investment. It involves energy infrastructure, access to networks, environmental conditions, availability of water or alternative cooling systems, tax benefits, permits, territorial impact and definition of responsibilities between private parties and the State.

The American experience shows that conflicts around AI are not resolved solely with promises of innovation. As artificial intelligence becomes more infrastructure intensive, projects need to explain more precisely where the energy will come fromwhat costs the communities will assume, what controls there will be and what specific benefits will remain in the country.

In Argentina, Stargate still appears more like a promise than a project with sufficient public details. This contrast makes the debate that is going through the United States today more relevant. The question is no longer just whether a country can attract OpenAI or other big tech companies.

It also matters under what conditionswith what transparency and with what information for the communities that could coexist with the physical infrastructure of artificial intelligence.