Every technology that mints a new fortune tends to make an enemy somewhere else. The railroads did it. The factory floor did it. Artificial intelligence (AI) is doing it now, and the backlash is showing up in places Wall Street doesn’t usually watch.For two years, the only AI story most investors cared about was the upside. Nvidia (NVDA) became the most valuable company on the planet. Data centers spread across rural counties. Microsoft (MSFT), Alphabet (GOOGL), Amazon (AMZN), and Meta (META) committed to spending sums that rival the budgets of small nations. If you own an index fund, a fat slice of your recent gains rode on that buildout.Outside the markets, the mood is colder. Towns are fighting the server farms landing next door. Workers are watching jobs disappear into software. Electricity bills are climbing in places where the grid now feeds data centers before it feeds homes.That anger has reached the one audience that can change how this story ends. Federal law enforcement has stopped treating AI opposition as a local zoning squabble. The FBI and the Department of Homeland Security (DHS) have begun tracking what they call “anti-tech extremism,” a new domestic threat category aimed at the people pushing back on the machines.What the federal documents actually revealThe shift sits inside more than 1,000 pages of unpublished government reports obtained through public records requests, according to Wired, which first reported the story.The material comes from DHS, the FBI, and the regional fusion centers that pass intelligence among federal, state, and local police.More Social Security: Social Security beneficiaries just got some shocking newsAARP warns Americans on major Social Security problemYoung Americans have a surprising plan for Social SecurityOne report from New York’s Intelligence and Counterterrorism Bureau warned that the “chaotic atmosphere” around AI over the next five years could feed protests, civil unrest, and “anti-tech violent extremist activity,” especially in big cities.Here’s the part civil liberties lawyers flagged. That brand-new phrase doesn’t appear in any public DHS or FBI guide to domestic extremism.The timing isn’t an accident. The reports follow President Donald Trump’s National Security Presidential Memo 7, which directs agencies to target organized political violence and a long list of “anti-American” beliefs.When I read through the documents and the coverage around them, the word that kept jumping out wasn’t “violence.” It was “broad.”Critics say the net is wide enough to catch peaceful data center protesters, AI skeptics, and people who simply don’t like where technology is headed. Intelligence reports have a long record of treating protest, or even strong opinions, as a precursor to violence, said Spencer Reynolds, a lawyer with the NAACP Legal Defense Fund.
The FBI and Homeland Security are now tracking “anti-tech extremism” as a domestic threat.Photo by Cravetiger on Getty Images
Why the AI backlash has turned violentThe government’s worry didn’t come from nowhere. The past few months delivered the kind of incidents that get security officials’ attention.In April, a 20-year-old man threw a Molotov cocktail at the San Francisco home of OpenAI chief executive Sam Altman, then threatened to burn down the company’s headquarters. “Thankfully, no one was hurt,” OpenAI said, according to CNBC. The suspect was carrying a manifesto that listed other AI executives.Days earlier, someone fired 13 shots at the home of an Indianapolis city councilman who had backed a data center project, leaving a note that read “No data centers.” The Washington Post described it as “a dark form of anti-technology politics” taking shape.The opposition runs far wider than a handful of violent actors, though. Hundreds of organizations across 42 states have formed to block data center construction, according to Data Center Watch, a project that tracks the fights.Here’s the scale of it, in plain numbers:More than 1,000 pages of unpublished DHS, FBI, and fusion center reports, according to WiredOpposition groups organizing across 42 states, Data Center Watch confirmedRoughly $700 billion in combined 2026 AI spending by Amazon, Microsoft, Alphabet, and Meta, CNBC reportedSpeakers removed or arrested at data center hearings in at least six states, according to WiredWhat the AI counterterrorismcrackdown means for your moneyIt’s tempting to file this under culture war and scroll past. That’s a mistake if you own stocks.The AI trade is the market right now. A handful of AI-driven names carry an outsized share of the S&P 500’s value, which means your 401(k) is more exposed to this fight than you probably realize.Related: Pope Leo XIV teams with Anthropic founder on bold AI warningIn my analysis, the signal for investors is simpler than the alarming language suggests. When Washington starts assigning counterterrorism resources to protect data centers and tech executives, it has picked a side. The buildout has federal cover now.For the AI names, that’s a tailwind. The fear that regulators might slam the brakes on AI looks overdone when the same government is guarding the infrastructure with terrorism statutes.The risk that won’t show up on a stock chart is social. Angry towns, climbing power bills, and workers who feel erased are how an industry quietly loses its license to operate. What to watch as the AI fight escalatesThe reports are warnings, not arrests. The FBI says it pursues people who commit or plan actual crimes, not those who hold opinions.But the gap between a protester and a “pre-operational” threat is exactly what civil liberties groups are worried about, and it’s the thing to watch.For your portfolio, the tell won’t be the next manifesto. It’ll be the boring stuff: local permits, power contracts, and how fast new data centers actually get approved. The violence makes the headlines. The zoning board moves the stock.The AI boom paid for a lot of gains this year. Whether it keeps its welcome, on Wall Street and on Main Street both, is the question that decides what your statement looks like next year.Related: AI is getting worse as Google and Anthropic nerf AI models and limit usage