An 84-year-old senator stood at Stanford last Friday and told a room full of future tech workers that America has “not a clue” what’s about to hit it. Then he proposed freezing the construction of AI data centers until Congress catches up.

Bernie Sanders isn’t wrong about the problem. He might be dangerously wrong about the solution.

The Stanford Bombshell

Sanders shared the stage with Congressman Ro Khanna — the guy who literally represents Silicon Valley — at an event titled “Who Controls the Future of AI: The Oligarchs or the People.” Subtlety wasn’t on the agenda.

The senator compared AI to every major technological revolution in history and said they all “pale in comparison.” He cited projections that AI and robotics could wipe out tens of millions of jobs in the coming decade. He mentioned a D.C. restaurant offering Valentine’s Day specials for people and their “AI buddies.” The audience laughed. Sanders didn’t.

“What is the long-term impact if we lose work as an important part of our lives?” he asked. “What do we do with our lives?”

These are the kinds of questions that rarely surface in policy debates dominated by GDP projections and quarterly earnings. Sanders is forcing them into the open — and that matters, even if his proposed fix doesn’t hold up.

A Moratorium Is a Sledgehammer for a Surgery Problem

The headline proposal: freeze AI data center expansion until Congress establishes guardrails. He plans to introduce legislation when he returns to Washington.

The logic isn’t crazy. Data centers are devouring electricity, straining power grids, and guzzling water for cooling. They’re the physical backbone of the AI revolution. Freeze the backbone, slow the revolution.

But here’s what actually happens if you freeze US data center construction: the builds move to China, the UAE, and anywhere else that wants them. You don’t stop AI development. You just ensure America isn’t leading it.

Even Khanna — Sanders’ ally and co-headliner — publicly disagrees. He’s pushing a “Singapore model” instead: smart regulation that mandates renewable energy and water efficiency rather than an outright ban. Steering beats stopping.

The Numbers Are Real (Even If They’re Debatable)

Sanders isn’t manufacturing anxiety. His Senate report estimated AI could replace up to 97 million US jobs over the next decade. The World Economic Forum projects 85 million displaced globally. A Gartner study found 37% of business leaders plan to replace workers with AI by end of 2026.

The public feels it. Pew Research found 64% of Americans think AI means fewer jobs. Only 17% see a positive impact.

But the counter-data exists too. The Bureau of Labor Statistics projects 3.1% employment growth over the next decade. Brookings found that of the 37 million US workers most exposed to AI automation, 26.5 million have strong “adaptive capacity” to transition into new roles.

History’s on the optimists’ side — every major technological revolution has destroyed jobs and created more. Automobiles killed the horse-and-buggy industry and spawned millions of new positions.

The difference this time? Speed. Previous transitions unfolded over decades. AI is moving in years.

The “New Gilded Age” Frame That Hits Different

Khanna described the current moment as a “new gilded age” run by tech billionaires who believe “they would have been heroic conquerors in a different era.”

“That’s just not my observation,” he clarified. “That’s what they tell me.”

Coming from the congressman who represents these exact people, that line carries weight. He outlined seven principles to prevent “oligarchic capture” of AI wealth, echoing JFK: “We must ask not what America can do for Silicon Valley, but what Silicon Valley must do for America.”

Sanders’ trip included helping launch a California ballot initiative for a one-time 5% tax on residents worth over $1 billion. Some billionaires are reportedly already fleeing the state over it.

The through-line is clear: Sanders sees the AI revolution as being driven by a handful of people with every incentive to maximize automation and minimize labor costs, with zero democratic accountability for the fallout.

Why This Isn’t Just Another Tech Panic

We’ve been here before. ATMs didn’t kill bank tellers. Spreadsheets didn’t kill accountants. Self-driving cars still haven’t replaced truckers.

But this time actually is different, for three reasons:

Generative AI hits different. It doesn’t just automate repetitive physical tasks. It handles creative, analytical, and communication work — the stuff knowledge workers thought made them irreplaceable. Legal briefs, marketing copy, medical image analysis, software engineering. The “safe” jobs aren’t safe anymore.

The pace is absurd. ChatGPT launched in November 2022. Three years later, AI generates photorealistic video, passes professional licensing exams, and writes production code autonomously. The capability curve isn’t flattening.

Power concentration is extreme. A handful of companies control the frontier models. The infrastructure requires billions. This isn’t the early internet, where anyone with a laptop could build something meaningful. The barriers to entry are enormous and growing.

What Should Actually Happen

Sanders is right that AI development is outpacing democratic institutions’ ability to respond. Congress is still figuring out how social media works.

But a data center moratorium is the equivalent of banning new factories during the Industrial Revolution. It wouldn’t stop the revolution — it would just guarantee you lose your seat at the table.

Better approaches exist:

  • Require renewable energy for new data centers — address the environmental costs without killing the industry
  • Mandate transparency in AI capabilities and limitations
  • Establish federal standards for AI in hiring and employment decisions
  • Build real transition programs for displaced workers, not the usual retraining theater
  • Create a digital dividend — if AI generates massive wealth, the distribution question is a legitimate political issue

The most urgent need is the simplest: making Congress actually understand what’s happening. You can’t regulate what you don’t comprehend.

The Real Win

Sanders’ moratorium almost certainly won’t pass. The political will isn’t there, and the tech lobby is powerful. But that might not be the point.

What Sanders is doing — more effectively than anyone else in Washington — is making AI policy a populist issue. He’s connecting it to the economic anxieties that millions of Americans already feel. And he’s doing it at Stanford, in the belly of the beast, refusing to be polite about it.

The AI revolution is coming whether we slow it down or not. The question Sanders is really asking isn’t whether we should stop it. It’s whether ordinary people will have any say in how it unfolds.

Given how quickly things are moving, that might be the most important political question of the decade.

Sources: The Guardian, VTDigger, SF Chronicle, Pew Research, Brookings