AI Is Making the Rich Richer at a Speed History Has Never Seen Before. Here’s the Math.

Two numbers. Hold them in your head at the same time.

Workers with AI skills earn 56% more than their peers without them. Workers whose jobs are being automated by AI are losing income with no clear path to replacement roles. Those two realities are happening simultaneously, in the same economy, right now.

This isn’t a prediction about the future. This is PwC’s 2025 Global AI Jobs Barometer describing the present.

Who Is Actually Getting Rich From AI

OpenAI is valued at $852 billion. Nvidia briefly became the most valuable company in human history on the back of AI chip demand. The handful of people who owned significant equity in the right companies at the right time have accumulated wealth at a pace that has no historical precedent.

Sam Altman’s net worth has increased by more than most countries’ GDP in the span of three years. Jensen Huang’s wealth grew faster than entire industries. This isn’t criticism, it’s math. The value creation is real. The concentration of who captures it is also real.

And below the billionaires, there’s a second tier doing extremely well. The engineers who can build AI systems. The product managers who can deploy them. The consultants who can explain them to executives. That 56% wage premium is compounding. Every year they earn more than their peers, they invest more, build more equity, widen the gap further.

Who Is Getting Left Behind

The World Economic Forum estimates AI and automation could displace 92 million jobs by 2030 while creating 170 million new ones. That sounds fine until you ask who gets the new jobs.

The 92 million displaced jobs are concentrated in routine cognitive work. Data entry. Customer service. Basic analysis. Legal document review. Financial report generation. These are jobs held by people with some education, moderate skills, and salaries in the $40,000 to $80,000 range. The middle.

The 170 million new jobs are concentrated in AI development, data analytics, cybersecurity, and clean energy infrastructure. These require advanced technical skills that take years to develop. A 45-year-old whose accounting job was automated can’t become a machine learning engineer in six months. The retraining pipelines don’t exist at anywhere near the scale needed.

The Wage Premium Is a Trap

The 56% AI wage premium sounds like an opportunity. Learn AI skills, earn more. Simple.

But wage premiums exist because supply is scarce. As more people develop AI skills, the premium compresses. The people who got in early, who got the premium for five years, used that income to build wealth, invest in the companies, own the assets. The people who arrive late get a smaller premium on top of a higher cost of entry.

This is how every technological transition works. The first wave gets rich. The second wave does okay. The third wave wonders what happened.

The Number That Keeps Me Up at Night

$650 billion. That’s what Amazon, Nvidia, Meta, and Microsoft are spending on AI infrastructure in a single year. Not revenue. Capital expenditure. Money going into chips, data centers, power infrastructure, cables, cooling systems.

That $650 billion gets extracted from somewhere. Some of it comes from revenue, some from debt, some from equity. But the operational costs get passed on. They come out of wages through hiring freezes. They come out of supplier margins through price pressure. They come out of communities where data centers replace manufacturing plants but employ a fraction of the workers.

The richest companies in history are spending unprecedented sums to build systems that will reduce their dependence on human labor. They’re doing this rationally. It makes financial sense. And the cumulative effect, playing out across millions of individual rational decisions, is the fastest redistribution of economic power in human history.

You’re either on the right side of that redistribution or you’re not. And the window to get on the right side is getting smaller every quarter.