If AI Steals Everyone’s Jobs, Who Will Buy the Products?

That question shows up under almost every post I write about artificial intelligence.

A hundred times a week, in a thousand different versions.

And it may be the smartest question in the entire AI debate.

So smart, in fact, that it was asked long before ChatGPT, long before Silicon Valley panic, and long before “AI disruption” became a daily headline.

Seventy years ago, in fact.

The 1954 answer

Cleveland, Ohio, 1954. Ford has just opened the first fully automated engine plant in history.

A company executive is giving Walter Reuther, the head of the American auto workers’ union, a tour.

Standing in front of machines that work on their own, he cracks a joke:

“Walter, how do you expect to collect union dues from these robots?”

Reuther doesn’t hesitate.

“And how do you expect to get them to buy cars?”

It’s hard to improve on that answer. In seventy years of papers, conferences, books, and panel discussions, nobody has come up with a sharper one.

And yet the robots at Ford are still there. Cars are still being sold. The system did not collapse.

Now the same question is back, only this time the scale is much bigger.

Why is the fear growing

Dario Amodei, CEO of Anthropic, has warned that AI could eliminate around 50% of entry-level white-collar jobs within five years, with unemployment potentially rising to 10% to 20%.

Jack Dorsey has cut thousands of jobs at Block while explicitly tying the move to AI. Oracle has also carried out major layoffs. In the first quarter of 2026 alone, the tech sector shed tens of thousands of jobs worldwide.

So the question is no longer theoretical.

It is starting to hit payrolls, hiring, and entire business models.

But the answer is not as simple as “the economy will collapse” or “we’ll adapt like always".

The truth is messier.

First: the real customer is companies

Most of the money in general-purpose AI does not come from individual consumers. It comes from businesses. Anthropic and OpenAI both rely heavily on enterprise demand.

That means the AI economy can keep growing even if ordinary people are under financial pressure.

One company buys models from another.

Another company buys cloud services.

Another buys chips.

The whole thing becomes a corporate chain reaction.

So even if workers lose jobs, AI companies can still make money by helping other businesses cut labour costs.

Second: wealth concentrates fast

When labour gets replaced by software, the value does not disappear. It moves upward.

It goes to the owners, the investors, the executives, and the people who control the systems.

That’s why luxury keeps booming even when the middle class feels squeezed. Hermès keeps selling bags. Ferrari keeps setting records. Private jets still have waiting lists.

The result is a split economy.

One side consumes at the top.

The other side is pushed downward or kept afloat by wage replacement, benefits, or state support.

Third: AI will make things cheaper

This part is easy to miss.

AI does not just destroy jobs. It also lowers the cost of making things.

That means cheaper software, cheaper content, cheaper services, and eventually cheaper physical goods too. Analysts increasingly describe AI as a deflationary or disinflationary force because it raises productivity so sharply.

In plain English: if production gets cheaper, prices can fall.

That matters because lower prices can partially offset lower wages.

A person with less income may still be able to buy more than before if the cost of living drops enough.

Fourth: the state will step in

This is the uncomfortable part.

If enough jobs disappear, governments will not just watch passively.

They will respond with some form of income support, redistribution, or public compensation for the losses caused by automation. Sam Altman has publicly discussed “universal basic wealth", a version of shared AI-generated prosperity. 

Call it what you want.

A basic income.

A subsidy.

A digital dividend.

The point is the same.

People will be given just enough to keep consuming.

Not because anyone is generous.

Because the system needs demand to keep moving.

The real split

This is where the future starts to bifurcate.

On one side, you have the owner class.

They own companies, brands, software, IP, property, and reputation. They live off returns, rents, and assets.

On the other side, you have the managed mass.

They live on support, consume cheap or free products, and remain economically dependent.

In between sits a shrinking middle class, slowly squeezed by automation, consolidation, and wage pressure.

That middle is not disappearing overnight.

It is being compressed over time.

The real question

So the original question is the wrong one.

“Who will buy the products?” is the question of someone who expects a salary and worries whether the system will still pay it.

The better question is the following:

“What will you own or sell in the new economy?”

That is the real fork in the road.

Because in the age of AI, survival is not just about getting a job.

It is about owning something that still matters when machines can do the work.

Reuther was right and wrong

Walter Reuther was right about one thing: robots do not buy cars.

But he was wrong about the bigger picture.

Cars still get bought.

Products still get sold.

The system changes shape, but it doesn’t stop.

Every few decades, capitalism reinvents itself, clears out the comfortable, and rewards the people who understood the shift early.

That is happening again now.

The question is not whether the market will survive.

It will.

The question is where you will stand when it finishes changing.