Meta steals a tactic from Tesla and builds data centers in tents
Meta is using tent-like rapid deployment structures to speed up and cut costs for AI data centers.
TechCrunch reports that Meta has built large tent-like “rapid deployment structures” near New Albany, Ohio, aiming to halve data center completion time. Cleanview’s Michael Thomas cited permits and satellite imagery showing multiple 125,000-square-foot structures built between April and June 2026. The setup, paired with modular gas turbines, highlights how AI infrastructure demand is pushing companies toward faster, cheaper, and more unconventional buildouts.
Meta may have found a way to reduce the enormous spending on AI data centers: building some of the data centers inside tents. According to TechCrunch, Meta has constructed six large tents near New Albany, Ohio—what Meta calls "rapid deployment structures"—with the aim of cutting data center completion time in half. This information comes from Michael Thomas, founder of Cleanview, which tracks data center deployment; after reviewing local permit filings and satellite imagery, he noted that Meta began constructing five tent-style facilities, each 125,000 square feet, between April and June 2026, and the satellite images he shared show that these structures are already complete. This approach is reminiscent of when Tesla, to accelerate Model 3 production, erected a tent-style production line in the parking lot of its Fremont factory; meanwhile, around 200MW of modular gas-turbine power supply near the site has also been described in reports as similar to a strategy often used by xAI. The article notes that the tents are expected to house AI chips potentially worth billions of dollars. The backdrop is that Meta is facing schedule pressure on the rollout of its AI models and developer APIs; the report mentions that the Wall Street Journal recently said Meta's latest model, Muse Spark, is complete, but the API that developers use to access the LLM from applications has been repeatedly delayed. Meta previously also stated that data centers and other capital expenditures could reach as high as $145 billion, which has given Wall Street pause, and Meta's stock has also fallen about 5% this year. As a result, placing AI chips into more quickly erected tent-style facilities has become one option for Meta to try to lower costs and speed up deployment in the AI infrastructure race. However, the article does not provide Meta's official response to this strategy; TechCrunch said it has reached out to Meta and will update upon receiving a reply.
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on TechCrunch AI →Summaries are AI-generated; the original article is authoritative.