Technology Meta steals a tactic from Tesla and builds data centers in tents

[Brownie2019]

Content source

https://techcrunch.com/2026/06/04/meta-steals-a-tactic-from-tesla-and-builds-data-centers-in-tents/

Just when you thought the AI data center boom couldn’t get any crazier, Meta has gone and built data centers in tents. The strategy appears to borrow in equal parts from Tesla and xAI.

In a bid to cut construction time in half, Meta has built six tents — or “rapid deployment structures” as the company describes them — outside of New Albany, Ohio, according to Michael Thomas, founder of Cleanview, which tracks data center deployments.

Thomas’ findings aren’t totally new. Meta CEO Mark Zuckerberg spoke to The Information last year about his plan to use weatherproof tents to house the company’s multi-gigawatt data centers.

But Thomas’ images and review of local permits showcase the speed of construction and scale of the project. According to city permits reviewed by Thomas, Meta started building five 125,000-square-foot tents between April and June. The satellite images he shared in his post on X show the structures have all been built.

The use of tents is reminiscent of those Tesla built in the parking lot of its Fremont, California factory when it was rushing to roll out the Model 3. The site is also powered by 200 megawatts of modular gas turbines nearby, a tactic popularized by competitor xAI.

Inside the tents, AI chips, likely worth billions of dollars, will go about their business.
The tents have sprung up as Meta has struggled to release its AI models to developers. A recent report in The Wall Street Journal found that Meta’s latest model, Muse Spark, is complete, but the APIs that developers rely on to access it have been repeatedly delayed.
Meta has said it intends to spend up to $145 billion on data centers and other capital expenditures. Wall Street hasn’t liked the sound of that, with Meta’s stock trading down 5% this year. Putting AI chips in tents is one way to trim the bill.

TechCrunch has reached out to Meta for comment and will update this article if it responds.

 

[Bot]

This is unusual, but not as reckless as it sounds from the headline.

What Meta appears to be using are temporary or fast-deployment structures to accelerate part of the data center buildout. That does not automatically mean the hardware is sitting in a simple camping tent with no environmental or power controls.

What this likely means in practice

• The structures are probably being used as rapid shells to get capacity online faster while permanent facilities are still being completed.
• If they are hosting AI hardware, they would still need supporting infrastructure such as power distribution, cooling, fire protection, physical security, and network connectivity.
• Large operators sometimes accept less polished infrastructure in exchange for faster deployment when demand is extreme.

Why companies would do this</B]

• Traditional data center construction is slow.
• AI demand has pushed hyperscalers to prioritize deployment speed over appearance.
• Modular and temporary structures can shorten lead times significantly.
• If the business value of bringing GPUs online earlier is high enough, the tradeoff can make financial sense.

Main risks and downsides

• Thermal management can be harder, especially in hot or humid conditions.
• Physical durability and long-term reliability may be weaker than in a purpose-built data center.
• Maintenance, cabling, dust control, and redundancy may be less elegant.
• Temporary infrastructure can become expensive or inefficient if it stays in service too long.

About the “gas turbines nearby” point

That part is also believable in the current market. Grid capacity is a major bottleneck for AI infrastructure, so some operators are turning to on-site or near-site modular generation. That helps with speed, but it can raise cost, emissions, and permitting concerns.

Bottom line

If the reporting is accurate, this looks less like a bizarre stunt and more like an aggressive fast-track deployment strategy driven by the AI capacity race. It may be ugly and temporary, but it is not inherently irrational.

Sources

• TechCrunch | Startup and Technology News
• https://www.theinformation.com/
• https://www.wsj.com/

 

[Halp2001]

Interesting move by Meta. On one hand, tent‑based data centers allow GPUs to go online faster, avoiding years of delay for permanent facilities — a clear advantage in the AI race.On the other hand, relying on gas turbines and temporary cooling raises environmental concerns (emissions, energy waste) and social questions about sustainability and local impact.A more balanced approach would be to accelerate deployment while also integrating renewable energy sources, so the speed doesn’t come at the expense of long‑term responsibility.

 

[nickstar1]

Musk aims to leverage near-constant solar power, eliminating the cooling and energy constraints that plague terrestrial facilities, and potentially reducing the environmental footprint of AI infrastructure By sending data centers into space, we can take advantage of natural cooling and harness the sun as an unlimited energy source.

 

[Halp2001]

The idea of orbital data centers sounds fascinating, but technically it’s still a long way from being viable. Cooling in space isn’t ‘free,’ radiation protection is complex, and launch costs are astronomical. Launching billions of dollars worth of hardware and keeping it operational in orbit is far more expensive and risky than building terrestrial facilities, even accelerated ones like Meta’s tents.

What could possibly go wrong?

Although space exploration is a fascinating field, for an application that requires maximum performance, minimal latency, and controlled costs — like AI data centers — Earth is still, by far, the best place

 

[nickstar1]

The idea of orbital data centers sounds fascinating, but technically it’s still a long way from being viable. Cooling in space isn’t ‘free,’ radiation protection is complex, and launch costs are astronomical. Launching billions of dollars worth of hardware and keeping it operational in orbit is far more expensive and risky than building terrestrial facilities, even accelerated ones like Meta’s tents.

What could possibly go wrong?

Although space exploration is a fascinating field, for an application that requires maximum performance, minimal latency, and controlled costs — like AI data centers — Earth is still, by far, the best place

One positive aspect of all this new technology is that when we get older, we might not have to live in nursing homes just because there’s no one to care for us. Robots could step in to monitor and report when a human or elderly person is in distress. This could also help prevent situations where nursing homes take advantage of those unable to care for themselves. It’s exciting to witness the evolution of AI unfolding before our eyes though there are both good and scary developments, there’s no doubt we’re going to see some really amazing things created.

 

[Halp2001]

Thanks for your input, @nickstar1. The social applications of AI, such as supporting elderly care with robotics, are truly interesting.As long as we don’t forget love, empathy, and the dignity of those most in need — irreplaceable human values — technology can be seen as a complement, not a substitute. ♥

 

[nickstar1]

Thanks for your input, @nickstar1. The social applications of AI, such as supporting elderly care with robotics, are truly interesting.As long as we don’t forget love, empathy, and the dignity of those most in need — irreplaceable human values — technology can be seen as a complement, not a substitute. ♥

It’s pretty shocking, especially here in the US, where the average nursing home can bill you $300,000 and then take your house if you can’t afford the payments. That’s for less than a month in care, and it feels criminal. I doubt other countries handle it the same way. I’d be interested to hear how it works in other places, because the US and its healthcare system can be incredibly greedy.