Somewhere in Virginia, a windowless building the size of an airport draws the power of a small city, and you’re paying for it. Residential electric bills are up forty percent since 2021. The same servers raising your rates are running deportation maps for ICE. The data center, the electric bill, the federal contract, and the surveillance state have become one machine, and almost no one has explained how the pieces connect.
A worker in Hong Kong thought he was on a video call with his company’s CFO and colleagues. He was not. Every face and every voice had been generated by AI, and by the time the illusion broke, $25.6 million was gone. This is the new cybersecurity threat: not hackers breaking into your machine, but attackers learning how to manipulate the human being sitting in front of it.
On February 27, 2026, the Pentagon blacklisted Anthropic for refusing to give the military unrestricted access to its AI. Hours later, OpenAI signed a deal on nearly identical terms. The media wants you to pick a hero and a villain. But the real story isn’t about which company is right. It’s about why two private companies were negotiating the rules of AI in warfare at all, and where Congress has been while the most powerful technology since nuclear energy gets handed to the most powerful military on earth with no binding laws governing how it’s used.
Beneath the apps and headlines sits a quieter reality: more of modern life now runs on three cloud platforms. Amazon, Microsoft, and Google didn’t “take over the internet” so much as win the race to build the world’s default plumbing, and their scale brings real benefits, from security to speed. The real worry isn’t whether these companies are good or bad. It’s how much everything else depends on them, and how hard it is to leave once your data, software, and workflows are stitched into the same stack.
AI data centers look like ordinary warehouses, but inside they function like nonstop factories for computation, turning electricity into heat at a scale that forces tough local tradeoffs. In fast-growing clusters, they can strain power grids, push expensive upgrades, and intensify fights over who pays, while cooling demands raise new questions about water use in places already under stress. The debate is no longer abstract. It is about infrastructure, rates, reliability, and whether the AI boom gets built in a way that communities can actually live with.
