Since our prior coverage flagged Meta's AI chatbot handing over Instagram accounts and the broader digital security mess earlier this week, three more significant Meta developments have landed on June 4 — revealing a company moving at dangerous speed.

Tents. Actual Tents. For Billion-Dollar AI Chips.

Meta is housing AI hardware in weatherproof tents outside New Albany, Ohio.

According to Michael Thomas, founder of data center tracking firm Cleanview, Meta built five 125,000-square-foot tent structures between April and June of this year. Satellite imagery and city permits reviewed by Thomas confirm the structures are already standing.

Each tent is essentially a temporary warehouse for AI chips worth billions of dollars. The site runs on 200 megawatts of modular gas turbines — off the main grid entirely. Meta CEO Mark Zuckerberg previewed this strategy in a conversation with The Information last year, but the speed and scale of execution is remarkable.

The playbook is borrowed from two other tech giants. The tent construction mirrors what Tesla did in the Fremont, California factory parking lot during its frantic Model 3 rollout. The off-grid gas turbine power mirrors what Elon Musk's xAI has done at its Memphis supercluster. Meta is essentially copy-pasting crisis-mode manufacturing tactics into permanent AI infrastructure.

Why? Because Meta has committed to spending tens of billions of dollars on data centers and capital expenditures, and conventional construction timelines are too slow for the AI arms race. Tents cut construction time roughly in half, according to TechCrunch.

Wall Street is NOT impressed with the overall spend. Meta's stock is down 5% year-to-date. Putting chips in tents may be a cost-trimming measure, but it also signals something: the company is in full panic-sprint mode.

Meta's latest AI model is reportedly complete according to The Wall Street Journal — but the developer APIs needed to actually use it have been repeatedly delayed. The company has billion-dollar chips sitting in tents, and the software to use them isn't ready.

Meta's Own Oversight Board Just Torched Its Account Ban System

On the same day the tent story broke, Meta's Oversight Board — the independent body Zuckerberg created and funds — released a scathing review of how Meta handles account bans.

The verdict: no due process, no transparency, no meaningful support.

The board launched its investigation earlier this year after reviewing a case involving threats of violence against a journalist. Meta was right to permanently ban that account, the board agreed. But digging into the broader system, the board found what it called "systemic human rights concerns."

Meta uses two separate enforcement tracks: one involving escalating strikes, another for instant permanent disabling of accounts for "egregious" violations. The board found the line between the two is not clearly defined or documented. Users don't know what puts them in which category.

Meta charges users for "Meta Verified" subscriptions — sold partly on the promise of 24/7 customer support. But when those verified users get their accounts disabled, they get zero meaningful assistance. Customers paid for help that vanishes when they need it most.

TechCrunch has been tracking this for years. Retired LA County firefighter Richard Pauwels had his account permanently banned with no specific post identified and no human review. Automated child sexual exploitation flags are sweeping up innocent users and destroying their accounts and businesses. There's no real appeals process.

This is a massive liability problem for Meta. Users have filed lawsuits, and more are in process.

The Oversight Board's funding was recently extended through 2028. Meta created this board. Meta funds this board. And the board is publicly saying Meta's enforcement system is broken. That's the company's own governance structure raising serious questions.

AI Creator Tools: The Good News Buried Under Everything Else

Amidst the chaos, Meta did announce something genuinely useful on June 4: a new AI creator assistant rolling out on Facebook in the U.S., Canada, and India.

The tool gives creators conversational access to their own performance data — answering questions like "When should I post?" or "What are people saying in my comments?" without requiring them to dig through dashboards. It can also suggest trending audio or content tied to cultural moments.

Over 500 million users on Facebook are now watching AI-translated videos weekly, Meta says. The company is also adding Arabic, Bahasa Indonesian, French, Thai, and Vietnamese to its AI translation feature for Reels.

This is Meta's direct play to keep creators from defecting to TikTok and YouTube. It's also a move to pull creators away from third-party tools like ChatGPT and keep them locked inside Meta's ecosystem. Smart business, honestly.

The Broader Picture

Meta is simultaneously: spending aggressively at breakneck speed (tents included), facing a legitimate governance crisis around account enforcement, and rolling out AI tools it's betting its future on — while the underlying developer APIs aren't even ready yet.

The tent story gets played as quirky tech-bro innovation. The Oversight Board story gets buried. The creator tools get a press release treatment. The reality is a company under enormous pressure making high-speed decisions with real consequences for real people.

For the millions of users who've had accounts wrongly banned — people who've lost businesses, income, and personal histories — no amount of AI creator tools matters. Fix the enforcement system first.

Zuckerberg is building the future in tents. The people who got locked out of their accounts are still standing in the cold.