xAI and Military AI: A Controversial Data Center

A Dispute Over a Data Center in Memphis

A data center located in Memphis, Tennessee, is at the center of a legal conflict that could have significant implications for U.S. national security. According to the digital and artificial intelligence officer of the Trump administration, the closure of this center would directly threaten the government's ability to protect the country. This dispute began in April when the NAACP filed a class-action lawsuit against xAI, the artificial intelligence company founded by Elon Musk. The complaint accuses xAI's Colossus 2 data center of violating the Clean Air Act. This center, which includes a power plant equipped with 27 gas turbines in Mississippi, is accused of operating without the required federal permits, generating significant pollution that endangers the health of nearby Black communities in Tennessee and Mississippi.

On June 15, the U.S. government filed a motion for the court to dismiss the case. Cameron Stanley, the digital and artificial intelligence officer at the Department of Defense, emphasized in the filing that the Memphis data center is essential for xAI to provide its AI services to its clients, including the U.S. government.

Grok AI: An Asset for National Security

The importance of artificial intelligence for national security is highlighted by the use of a special version of Grok, named Grok Gov Model, by the U.S. government. This technology is integrated into the Maven Smart System, a military AI system developed by defense giant Palantir, used by the United States and NATO. This system is crucial for functions such as targeting, intelligence, preparation, and recruitment, according to Stanley.

The Department of Defense has signed agreements with eight major AI companies, but Grok stands out as one of three providers capable of supporting critical operations, including during covert missions. Stanley pointed out that Grok has unique capabilities that make it indispensable for national security. He cited Operation Epic Fury, a bombing campaign conducted by the U.S. and Israel against Iran, as an example of Grok's effectiveness. During this operation, American defense systems incorporating Grok enabled the deployment of over 2,000 munitions on as many targets in just 96 hours, demonstrating increased operational efficiency thanks to the Grok Gov Model.

Community Reactions and Environmental Issues

The proliferation of data centers has sparked strong reactions from local communities, and the NAACP's case raises crucial questions: Are data centers essential for national security if AI is used in military conflicts? AI is already widely integrated into military and defense industries. In May, the chief technology officer of the Department of Defense reported a 1,775% increase in AI usage compared to the previous year. In January, Secretary of Defense Pete Hegseth stated that the U.S. would become a military force prioritizing AI across all domains.

While AI is used for commercial tasks, defense officials believe it will enhance surveillance, targeting, and autonomous weapons. Government spending reflects these priorities, with a budget of $13.4 billion allocated for autonomous systems by 2026.

However, questions remain about the role that these new AI technologies and the data centers supporting them should play in military operations.

Conflict Between Environmental Legislation and National Security

In response to Stanley's statement, the NAACP reaffirmed its intention to continue defending democracy against what it sees as harassment and authoritarianism. Abre' Conner, director of environmental and climate justice at the NAACP, emphasized the importance of not allowing polluting industries to thrive at the expense of the health of Black communities, especially at a time when the ultra-wealthy seem to be protected by certain government entities.

Data centers have a notable environmental impact on surrounding communities. They require vast amounts of water to cool servers, consume large amounts of electricity, and their discharges can contaminate nearby neighborhoods. Local communities have often contested the operating permits of these centers on environmental grounds.

Government Intervention and Its Consequences

The Environmental Protection Agency and the U.S. government are often involved in environmental litigation, but the government's intervention in this case is unusual, according to Vincent Joralemon, a law professor at the University of California, Berkeley. "What is most unusual here is that the government is supporting a private polluter to dismiss a citizen action," Joralemon explains. "This is the first time the federal government has intervened in a citizen action under the Clean Air Act against a private company to seek dismissal of the case."

The Clean Air Act establishes federal standards for emissions of hazardous air pollutants from stationary and mobile sources. Large projects, such as data centers, require permits to operate, which includes implementing the "best available pollution controls." The NAACP argues that xAI does not have these permits.

AI and Its Role in the Military Domain

The adoption of new technologies by the military, from landmines to nuclear weapons to drones, has always raised concerns and intensified the stakes of war. The arrival of AI has amplified this debate in recent years.

Experts have expressed concerns when the Pentagon had a public dispute with Anthropic after the AI developer refused to allow the use of its AI Claude for surveillance and autonomous weapons. President Trump criticized the company, labeling it "woke," and since then, the Department of Defense has shifted two-thirds of its AI activities to other models.

Government AI can be used "from the back office to the battlefield," according to Michael Horowitz, a political science professor at the University of Pennsylvania and former deputy assistant secretary of defense. "There are military uses of AI for logistics or human resources, similar to those of any American business, up to potential uses closer to the battlefield to help commanders identify potential targets more quickly and accurately," Horowitz explains.

AI is already integrated into some government activities. Defense employees using the Maven Smart System process nearly 2 billion tokens each day, according to Stanley. Tokens are a unit of measurement for AI processing; for example, having a chatbot write a one-paragraph email might require 400 tokens.

Weapon targeting systems are an area where AI integration is being considered. While Grok is one of three AI models used for this work, it is unique in the industry. Grok is known for having less strict safeguards than those applied by other AI companies.

Stanley did not specify whether Grok or other AIs were used in the targeting and deployment of missiles during Operation Epic Fury, but an AI-powered version of that future is not out of the realm of possibility. These systems must have strict safety safeguards. We have already seen how fatal it can be when these operations fail.

In March, shortly after the Pentagon's dispute with Anthropic, the U.S. launched an airstrike against an Iranian elementary school, killing 156 people, including 120 children. In the days following the strike, concerns were raised about the use of AI in the mission.