AI Is Swallowing Trillions. What's Left for Everyone Else?

A Brookings Institution study protects $10.3 trillion in AI data investment through 2032.

That's in the next, let's see, the next six years.

Six years, $10.3 trillion.

This exceeds prior build-outs as a share of GDP.

I'm asking questions this week.

Is this a good thing? It seems that the current data center build-out is consuming more than just power and water.

It's consuming capital.

So what does that mean for federal, state, and local governments that need to borrow and fund projects? What does it mean for Wall Street? More importantly, what does it mean for Main Street? Let me tell you something right now, and I believe this with my whole heart.

This is free.

I'm not saying that I've got a crystal ball, but I've lived long enough to see a bubble and know a bubble when I see it.

And I'm hearing the signs.

Here are the signs.

Number one, excessive investment, and $10.3 trillion is excessive.

Um, the rise of way cheaper substitutes.

China has developed what they call open source AI models that cost a fraction.

And we're not even talking billions, barely millions, and, uh, single millions, not hundreds of millions.

They're building their data centers to be sustainable energy, and they're building them smaller.

The problem I've got is that this money is far and above what these companies in big data can generate on their own.

So they're coming to the market, and it's like a giant vacuum cleaner.

Because there's demand for debt, investors are gonna want a higher percentage.

That is for everybody.

And what are you gonna fund? Are you gonna fund a $2 trillion project as opposed to you needing to borrow a quarter of a million dollars for your business? And the fact that because they borrowed that and hoovered up, uh, the capital, making capital scarce now for everybody else and raising the price that we all have to pay to borrow money to run businesses, finance inventory, or expand, it's gonna impact everybody.

Cities and towns very often go to bond referendums.

They borrow money for infrastructure projects, sewer treatments, schools, uh, civic buildings, roads, highways, and they spread that out over 20, 30 years or more.

Again, they're in the same market.

A Brookings Institution study protects $10.3 trillion in AI data investment through 2032.

That's in the next, let's see, the next six years.

Six years, $10.3 trillion.

This exceeds prior build-outs as a share of GDP.

I'm asking questions this week.

Is this a good thing? It seems that the current data center build-out is consuming more than just power and water.

It's consuming capital.

So what does that mean for federal, state, and local governments that need to borrow and fund projects? What does it mean for Wall Street? More importantly, what does it mean for Main Street? Let me tell you something right now, and I believe this with my whole heart.

This is free.

I'm not saying that I've got a crystal ball, but I've lived long enough to see a bubble and know a bubble when I see it.

And I'm hearing the signs.

Here are the signs.

Number one, excessive investment, and $10.3 trillion is excessive.

Um, the rise of way cheaper substitutes.

China has developed what they call open source AI models that cost a fraction.

And we're not even talking billions, barely millions, and, uh, single millions, not hundreds of millions.

They're building their data centers to be sustainable energy, and they're building them smaller.

The problem I've got is that this money is far and above what these companies in big data can generate on their own.

So they're coming to the market, and it's like a giant vacuum cleaner.

Because there's demand for debt, investors are gonna want a higher percentage.

That is for everybody.

And what are you gonna fund? Are you gonna fund a $2 trillion project as opposed to you needing to borrow a quarter of a million dollars for your business? And the fact that because they borrowed that and hoovered up, uh, the capital, making capital scarce now for everybody else and raising the price that we all have to pay to borrow money to run businesses, finance inventory, or expand, it's gonna impact everybody.

Cities and towns very often go to bond referendums.

They borrow money for infrastructure projects, sewer treatments, schools, uh, civic buildings, roads, highways, and they spread that out over 20, 30 years or more.

Again, they're in the same market.