Data Centers in Permit Crisis: Google, SpaceXAI and $6.1M

Are data centers now approved at midnight?

Two separate data center stories this week raised the same question: why are these facilities being built with such haste and so little transparency?

The Hermantown City Council approved six agreements for Google's data center at midnight after a four-hour public hearing. But it's not over: three lawsuits are ongoing and the 500 MW power deal remains unresolved. It's a concrete example of how hard local governments are struggling against Big Tech. Read the story.

In the same week, SpaceXAI faced penalties of up to $6.1 million for building three data center facilities in Mississippi without the required stormwater drainage permits. That shows permits aren't just bureaucracy—they're a serious cost. The 'build first, permit later' era seems to be closing.

Google's data center in Finland, where it cut down 400,000 trees and triggered a criminal investigation, completes the picture. Data centers are no longer just about electricity and water; they're about law and reputation. Finland investigation.

AI writes code, but who secures it?

This week brought two opposite poles on AI coding tools: one set a record, the other raised a security alarm.

Five Qwen3.8-27B agents beat Fable 5 on LiveCodeBench Hard, scoring 9.8 to 8.0. But the test is small and contamination is possible, so the result is exciting but not definitive. Qwen agents story.

Meanwhile, Veracode's test found that 38% of AI-generated code fails basic security checks. Models are writing code fast, but one in three lines still carries a vulnerability. Veracode test.

Anthropic dropping its coding interview is another angle: Claude writes better code than candidates, so the interview is changing. That could signal a major shift in software hiring. Anthropic interview decision.

The human side: Gen Z won't let resumes look AI-written

Tech news isn't only about code, data centers and robots. This week, how young job seekers are strategizing in the AI era also stood out.

According to a Greenhouse survey of 1,100 applicants, Gen Z is editing resumes to avoid looking AI-generated. And 81% of applications are rejected without anyone speaking to the candidate. That shows both the pressure AI puts on hiring and how companies still view human effort. Gen Z resume story.

Boston Dynamics approving a $2 million Spot robot dog delivery to ICE despite a petition from 300 employees raises questions about the ethical limits of technology. Employees' voices aren't being heard; management is imposing its own decision.

Are air-gapped computers no longer safe?

The RAMBO research showed that even air-gapped computers can leak data via radio signals from their RAM.

Researchers demonstrated RAMBO, which turns an air-gapped computer's RAM into a radio transmitter. No Wi-Fi, USB or camera is needed for the leak. That means even high-security systems can no longer be protected by physical isolation alone. RAMBO story.

In the same week, AO3's 12-hour maintenance turned into a four-week outage, reminding us how fragile digital infrastructure can be. AO3 outage.

The week's standout numbers

The week's most striking figures show how big an economic and legal force data centers have become. The table below gathers the numbers that stood out from this week's stories.

What it means for you

For US readers, the Hermantown vote and SpaceXAI's penalty are local stories with national implications: permit fights over data centers are now happening in your backyard, and the outcomes affect power bills and local budgets. In the UK and EU, similar battles are playing out under stricter environmental rules, so expect more midnight votes and lawsuits.

If you write or review code, the 38% failure rate is a direct warning: AI assistants speed things up but don't replace security testing. And if you're job hunting, the Gen Z resume trend suggests authenticity now beats polish—at least until a human actually reads your application.

Next week, eyes will be on Google's Hermantown lawsuits and SpaceXAI's penalty process. AI code security looks set to stay on the software world's agenda. For tech teams everywhere, these stories show it's time to take permits and security seriously.