The Most Dangerous Demographic in History Is About to Get a Lot Bigger

Educated Young Men + No Economic Path = Explosion

Peter H. Diamandis
Jun 09, 2026

Every major revolution in modern history shares a common ingredient. Not ideology. Not technology. Not even poverty. The ingredient is young men, usually between 18 and 28, who are educated enough to know what they’re missing and blocked from any path to get it.

The Arab Spring. The French Revolution. The rise of Nazi Germany. The Iranian Revolution. The Bolsheviks. The pattern repeats with eerie consistency across centuries, continents, and political systems. And it may repeat again in the near future… driven not by famine or war, but by artificial intelligence.

THE PATTERN

Academics call it “Youth Bulge Theory.” The research, led by scholars like Jack Goldstone, Gunnar Heinsohn, and Henrik Urdal, is unambiguous: between 1970 and 1999, 80% of civil conflicts occurred in nations where more than 60% of the population was under 30. The formula is simple. When a society produces a large cohort of young men and fails to provide them with economic purpose, marriage prospects, and social status, instability follows. Not sometimes. Almost always.

The critical detail is that revolutions are not led by the poorest and least educated.

“The revolutions are led by the educated-but-blocked: young people with enough knowledge to understand the system and enough frustration to want to tear it down.”

In Tunisia in 2010, the man who lit the match was Mohamed Bouazizi, a 26-year-old with a college degree who couldn’t find work and was reduced to selling fruit from a cart. When police confiscated his cart, he set himself on fire. Within months, governments fell across the Middle East and North Africa. Youth unemployment in the region was 25-30%. The demographics were identical in every country that erupted: massive cohorts of men in their 20s with degrees and no jobs.

In Weimar Germany, the brownshirts weren’t recruited from the illiterate poor. They came from the educated lower-middle class, young men who’d been trained for careers in a modern industrial economy and then watched hyperinflation and the Depression destroy their prospects. Hitler’s core base was men under 30 who felt the system had cheated them out of the life they’d been promised.

The Iranian Revolution in 1979 followed the same script. The Shah had massively expanded university education, producing hundreds of thousands of graduates. The economy couldn’t absorb them. Youth unemployment for college graduates hit 30-40%. Khomeini didn’t create the anger. He organized it. The Revolutionary Guard was built from young men aged 18 to 25.

In 1848, revolutions swept across France, Germany, Austria, Hungary, and Italy simultaneously. The common thread: a generation of university-educated young men who found the old aristocratic order had no room for them. In Russia in 1917, young soldiers returning from World War I found no jobs, no food, no future. Lenin channeled their rage. Even ISIS, in the 2010s, recruited heavily from young men aged 18-28 in regions with 40%+ youth unemployment, many of whom had at least some college education.

The pattern holds across centuries: educated young men plus economic exclusion equals instability. Every time.

WHY AI MAKES THIS DIFFERENT

I’m typically the eternal optimist, but I do look at the data. Here’s what concerns me. Every previous wave of technological disruption displaced workers over decades. The industrial revolution took a full generation to restructure labor markets. The transition from agriculture to manufacturing played out over 50 years. Even the internet revolution, which moved fast by historical standards, gave workers roughly 15-20 years to adapt.

AI is compressing that timeline to months.

This week as we reported on Moonshots, Anthropic published data showing that more than 80% of the code merged into their codebase is now written by Claude. Their engineers are shipping 8x as much code per quarter as they were two years ago. On the same day, OpenAI’s head of reinforcement learning told an interviewer they’d “turn AI on AI itself” within six months. The models aren’t just doing work faster. They’re learning to do the work of making themselves better.

A 24-year-old graduating this spring with a computer science degree, $150,000 in student debt, and a plan to become a software engineer is walking into a labor market where AI can already do 76% of open-ended coding tasks successfully, up from 26% just six months ago. That number is not going down.

And software engineering is the canary. Legal research, financial analysis, medical diagnostics, content creation, customer support, consulting, translation, accounting: the list of knowledge-work categories where AI performs at or above entry-level human capability is growing every quarter.

The people this hits first and hardest are not manual laborers. They’re the exact demographic that has driven every revolution in modern history: educated young people who did everything they were told, took on debt to get degrees, and are now discovering that the economy has no use for them.

THE MISSING SAFETY VALVES

Previous generations had escape routes. In 19th-century Europe, disenfranchised young men could emigrate to the Americas. In post-Civil War America, they could go west. In the mid-20th century, expanding government and military provided employment floors. None of those valves exist at the same scale today.

There’s no frontier to absorb millions of displaced knowledge workers. Military forces are shrinking, not growing. Government employment is contracting in most Western democracies. And the gig economy, which once served as a buffer, is itself being automated.

Historically, the other safety valve was family formation. Young men with jobs get married, have children, buy homes, and acquire stakes in social stability. Young men without jobs don’t. In every revolution I’ve cited, the inability of young men to marry and start families was a compounding accelerator of rage. Marriage rates in the US for men under 30 are already at historic lows. AI-driven job displacement will push them lower.

One more factor that has no historical precedent: this generation of potentially disenfranchised young men is the most networked in human history. The Arab Spring proved that social media enables radicalization and coordination faster than any government can respond. The tools available now make 2011-era Twitter look primitive.

THE COUNTERARGUMENT AND WHY I’M STILL CONCERNED

The optimist’s case, and I am generally an optimist, is that AI will create new jobs faster than it destroys old ones. That has been true of every previous technological revolution. The automobile eliminated horse-related jobs and created millions of new ones. The internet destroyed retail jobs and created the entire digital economy.

I do believe that will happen again. AI will generate new categories of work we can’t yet imagine. But the timing matters enormously. If the destruction comes in 2-3 years and the creation takes 10-15, you have a decade-long window where tens of millions of young people in developed economies have no economic pathway. A decade is more than enough time for the pattern to repeat.

Jack Goldstone’s research shows that revolutionary conditions require three elements arriving simultaneously: elite overproduction (too many educated people for too few positions), fiscal crisis (governments unable to fund social programs), and mass mobilization potential (large cohorts of idle young people with the tools to organize). AI could deliver all three within the same compressed timeframe.

WHAT WE SHOULD DO ABOUT IT

This is not a prediction of inevitability. It’s a risk assessment. And the response needs to be proportional to the risk.

First, we need to radically restructure education. Four-year degrees that cost six figures and train people for jobs that AI can do are worse than useless. They’re incendiary. Education needs to move toward human skills that complement AI: judgment, leadership, creativity, physical-world expertise, and entrepreneurship.

Second, we need to accelerate new job creation, not just wait for it to emerge organically. Government policy, corporate investment, and entrepreneurial energy should be directed at identifying and scaling the new categories of work that AI enables, not just the AI itself.

Third, we need honest public conversation about timelines – there is NOT enough public discourse going on regarding this potential outcome. The worst possible outcome is to tell a generation of young people that everything will be fine, hand them diplomas they borrowed $150,000 to earn, and let them discover the truth on their own. That’s how you get revolutions.

The pattern is clear. The demographics are in place. The technology is accelerating. The question is whether we’re smart enough to learn from 400 years of history, or whether we repeat it.

PLEASE enter this conversation. Ideas welcome. On my end I’ve launched two XPRIZE intended to help address this outcome.

First, the Future Vision XPRIZE ( www.FutureVisionXPRIZE.com) intended to help steer the future, but creating film and video content showing positive human-AI futures.

Second, the “Build with Gemini XPRIZE” ( www.GeminiXPRIZE.com) to help inspire and teach entrepreneurs to build their own future (give them agency) rather than wait for a job to materialize.