How fear of artificial intelligence grew into real arsons and Molotov cocktails.
In 1980s sci-fi films, machines usually rose up against humanity — but in reality it went the other way around. US federal agencies — the FBI and DHS (Department of Homeland Security) — now have a brand new line of work.
They officially coined the term "anti-tech extremism." Law enforcement uses it to describe people who are taking their hatred of the AI industry out of online chat rooms and into actual physical violence.
Ideologically, this movement is a genuine patchwork. The same category covers anarchists, eco-radicals, people with personal grievances against large corporations, and those who are terrified of being replaced by machines. What ties them together is the conviction that AI is the defining enemy of the next decade — and a willingness to prove it by extreme means.
Here's what anti-tech extremists are actually doing (per FBI investigations):
Attacking data centers. Server farms have become the universal target. In November 2025, anarchists burned down an Equinix data center construction site near Paris; in January 2026, a German far-left group torched a facility near a Tesla factory, hiding behind environmental rhetoric.
Hunting down executives. In April 2026, a Molotov cocktail was thrown at the San Francisco home of Sam Altman (CEO of OpenAI). The suspect, Daniel Moreno-Gama, drove in from Texas carrying a manifesto that explicitly named AI leaders as primary targets.
Threatening officials. According to the Soufan Center, the most violent online threats right now are directed at local government officials whose only offense was approving permits for new server construction.
Public anxiety
Gallup data shows 61% of Americans believe AI will eliminate more jobs than it creates, and a Pew survey finds 57% see the technology as a high risk to society. And the trend has gone global: protests against data centers are happening in Spain, the Netherlands, Chile, Malaysia, and Finland.
AI companies used to treat theft of their digital models and cyberattacks as their primary risks. Now the scattered incidents have solidified into a pattern: the industry is being forced to harden construction site perimeters, hide executives' home addresses, and hire security for GPU warehouses.