DOJ Sides With xAI Over Memphis Turbines

The Department of Justice has thrown its weight behind xAI in a fight over the dozens of unpermitted natural gas turbines powering Elon Musk’s data centers near Memphis. According to TechCrunch AI, the DOJ filed a memorandum on Monday opposing a lawsuit from the NAACP that seeks to shut down the turbines, arguing that doing so would harm “American national, economic, and energy security.” The federal government’s reasoning leans on an unusual claim: that Grok helps run “mission-critical” military operations.

What happened

Here’s the situation in brief:

  • The NAACP, represented by the Southern Environmental Law Center, sued xAI in April to stop its use of “mobile” gas turbines at the Colossus and Colossus 2 data centers.
  • The DOJ sided with xAI, calling the turbines a matter of national security.
  • In its filing, the Justice Department said Grok is one of four AI models supporting Department of War operations, including recent strikes in Iran.
  • xAI has kept adding turbines despite the legal pressure. The count now sits at 57, more than double what it was a year ago.

The regulatory loophole

xAI’s defense rests on a technicality. Because the turbines stay mounted on trailers, the company claims they’re “mobile” and exempt from Mississippi air pollution rules for one year. The Southern Environmental Law Center disagrees. It argues federal law treats trailer-mounted turbines as stationary sources once they’re parked and running, which would make them subject to regulation.

That distinction is the whole ballgame. If a court rules the turbines are stationary, xAI would need permits it doesn’t have for a setup it’s already scaled to 57 units.

Why it matters

What stands out here is the framing. The federal government isn’t arguing the turbines are clean or properly permitted. It’s arguing that AI compute now counts as critical national infrastructure, and that environmental enforcement should bend around it. That’s a notable shift. Tying a commercial AI lab’s power supply to active military operations gives xAI a shield few companies could invoke.

The stakes for local residents are concrete. TechCrunch AI reports that the Memphis region is already one of the most polluted in the country, and air quality has worsened since the data centers came online. As the turbine count doubled, so did three major pollutants:

  • PM2.5 – linked to stroke, cardiovascular disease, and Alzheimer’s
  • Formaldehyde – raises cancer risk
  • Oxides of nitrogen (NOx) – tied to asthma and heart disease

This is the tension the whole AI buildout keeps running into. Training and running frontier models takes enormous power, and the grid often can’t deliver it fast enough. Companies are turning to on-site gas generation to fill the gap, which puts compute growth on a collision course with local air quality and permitting law.

What comes next

The turbine fleet isn’t shrinking. xAI is now a division of SpaceX, and SpaceX’s IPO filing spells out the plan: another $2.8 billion in gas turbines over the next three years to power its AI data centers. At least $2 billion of that is earmarked for the same “mobile” units now under legal challenge.

So this case is bigger than 57 trailers in Memphis. It’s an early test of whether the “mobile turbine” loophole holds up in court, and whether the national-security argument becomes a standard play for AI firms racing to power their data centers. If xAI wins, expect other companies to study the playbook closely.

A few things to watch:

  • The court’s ruling on “stationary” sources. This decides whether the loophole survives.
  • How far the national-security argument stretches. A win here could set precedent for other AI infrastructure fights.
  • Local health data. Continued pollution increases will keep pressure on regulators and could draw more legal challenges.

The core question is whether AI’s power demands get treated as a special case that overrides environmental rules, or whether the same permitting laws apply to everyone. The Memphis ruling will be an early signal. Full details are available at the original TechCrunch AI report.

Scroll to Top