The Economist does not do panic. So when its leader column this week declares that "America's free-wheeling treatment of AI looks as if it is coming to an end," you should probably pay attention.

The thesis is simple and devastating: the Trump administration built its entire AI strategy around deregulation and speed. Then Anthropic released Claude Mythos on April 7th, and the people who were most worried about overregulation suddenly realized they should have been worried about underregulation.

The Economist calls Mythos "so startlingly good at finding software vulnerabilities that, in the wrong hands, it would threaten critical infrastructure, from banks to hospitals." That is The Economist being restrained. The Financial Times, in a companion piece, reports that senior international financial officials have warned the latest AI models "could threaten the world banking system."

The two-front problem

What makes this moment different from every other "AI needs regulation" hand-wringing piece is that both sides of the political equation are now aligned on the problem, for completely different reasons.

National security hawks in the Trump administration are rattled because Mythos proves AI can breach critical infrastructure. American voters are angry because AI data centers are devouring their towns, their electricity, and their water. The Economist puts it bluntly: "A laissez-faire approach is no longer politically tenable or strategically wise."

This is not left versus right. This is reality arriving late to the party.

The hyperscale backlash

Mother Jones, in a sprawling new investigation, documents what this AI building spree actually looks like on the ground. Meta's planned Hyperion campus in Louisiana covers 5.7 square miles of farmland and would consume enough electricity to power New Orleans three times over. Zuckerberg helpfully posted an illustration showing the site overlaid on Manhattan, stretching from Harlem to SoHo. Because nothing says "for the people" like paving over a farming hamlet to power chatbots.

The reporting reveals a speculative frenzy. Third-party agents stalk bean fields on behalf of anonymous buyers. Utilities keep coal plants online. The White House slashes nuclear safety regulations. Pecan groves become RV lots. Homes become parking lots. Energy Secretary Chris Wright calls it "Manhattan Project 2." Mother Jones notes that data centers have replaced megayachts as the preferred theater of oligarchic status signaling.

Total AI infrastructure expenditure amounted to about a quarter of all US GDP growth in the first half of 2025. The largest companies collectively spent $400 billion on construction projects, many of which will not be fully operational for years.

The convergence

Here is what nobody is saying out loud: the national security panic and the voter backlash are about to merge. The same technology that can hack banks can also drink the water supply of a parish in Louisiana. The same companies asking Washington for deregulation are the ones pushing coal plants back online while the White House rewrites nuclear safety rules.

The AI industry spent the last three years telling anyone who would listen that regulation was the real threat. That China would win if America slowed down. That guardrails were overreaction.

Then one model proved the guardrails were never strong enough to begin with.

The Economist is right. The free-wheeling era is over. What comes next is the hard part: building regulation that is fast enough to matter and smart enough not to hand the race to Beijing. Given that Congress introduced 47 AI bills and forgot to define what AI actually is, I would not hold my breath.

Sources: The Economist (leader column, April 16), Mother Jones (May/June 2026 issue), Financial Times.