• TalentGenius
  • Posts
  • When AI’s physical footprint starts to hurt

When AI’s physical footprint starts to hurt

AI’s boom isn’t virtual—it’s draining power grids and reshaping communities.

Author

Malcolm Frank, CEO
September 22, 2025

When AI’s physical footprint starts to hurt

Last week Microsoft announced a second “supercomputer campus” in Wisconsin, pushing their AI tab around Racine above $7 billion. The press release frames it as progress; the reality looks more like an arms race, as this follows similar announcements from AWS (in Ohio and Pennsylvania), Google (in Oklahoma, Virginia and Indiana) and Meta (in Louisiana).

These aren’t data centers the way we once imagined them, with rooms of servers quietly humming. They’re industrial-scale furnaces of computation: hundreds of thousands of GPUs, exotic cooling rigs, power demands that rival small nations.

Here’s what most of the public misses: the cloud isn’t in the sky, it’s in your backyard—draining the grid, heating rivers, bending public infrastructure to its will. 

America’s data centers already pull down about 176 terawatt-hours a year. By 2030, that figure could easily more than double, sucking up nearly 10% of the country’s electricity. Put another way: the AI build-out is the energy equivalent of putting 18 million new Cadillac Escalades on our highways.

So, feeling guilty about a 16 MPG SUV? That will soon be a lot greener than AI.

Microsoft to spend $4 billion on second Wisconsin data center

Follows $3.3 billion spent on their first data center in Racine, will house hundreds of thousands of Nvidia chips, and will require more than 900 megawatts of power.  

Is enterprise automation starting to surpass augmentation?

New data from Anthropic’s Economic Index report shows that it is.

Where are the jobs for Gen Z?

The number of entry-level jobs has dropped by 29 percent between January 2024 and June 2025. Where should 20-somethings focus?

The perils of using AI to replace entry-level jobs

HBR article argues these roles are crucial for developing future leaders, fostering innovation, enriching organizational culture, and protecting society. Instead of cutting these jobs, companies should redesign them to leverage AI.

AI agents are getting ready to handle your whole financial life

“If you want to scale financial advice to 50 million clients, you’re not going to find enough advisers to do that…but AI is going to allow you to scale financial advice to a far larger population that may be looking for it.”

“I regret to inform you, Meta’s new smart glasses are the best I’ve ever tried.”

The new $800 Meta Ray-Ban Display glasses have a screen that can show directions, pictures, translations, and more.

Duolingo CEO: AI makes my employees “four or five times”as productive

AI’s automation capabilities help Duolingo’s engineers produce language, math, music and chess lessons more quickly, CEO von Ahn said. 

The internet made us stupid; AI promises to make it worse

The significant negative correlation between frequent AI tool usage and critical thinking abilities.

AI is learning to predict the future—and beating humans at it

At the recent Metacalculus forecasting conference, an AI platform scored 2x better than the top human futurists. “It’s actually kind of mind blowing.”

Case study: Moody’s cut credit memo prep from 40 hours to two minutes with AI agents

Rather than simply making existing processes faster, agents fundamentally change how work gets done by handling multi-step workflows that previously required significant human coordination and analysis.

The two-step comes to jobs platforms: Fiverr to cut 30 percent of workforce in AI pivot

The online gig platform pushed AI on their freelance marketplace — now AI is coming after company employees too.

Agent of the week: Zentester

Streamline testing with intelligent, adaptive test coverage.

Oxford gives students access to AI platform

Oxford has become the first UK institution of its kind to offer access to the education version of ChatGPT to all of its students and staff.

OpenAI’s research on AI models deliberately lying is wild 

AI’s “scheming:” a practice in which a model “behaves one way on the surface while hiding its true goals.”

How successful sales teams are embracing agentic AI

Harvard: Agentic AI is revolutionizing sales by enabling autonomous personal agents to work alongside human sales reps, identifying, nurturing, and closing deals across channels.

And one more thing…

Which diseases will you have in 20 years? This AI accurately predicts your risks

Delphi-2M analyses a person’s medical records and lifestyle to provide risk estimates for more than 1,000 diseases.

Reply

or to participate.