Where the Market’s heading next

Dear Reader,

This is Dylan Jovine with Behind the Markets.

Happy Wednesday.

Today is Wednesday, August 5th.

I’ve been getting a lot of great questions from members lately — the same ones keep coming up again and again — so today I want to answer four of them directly.

If you have a question, drop us an email or leave a comment and I’ll get to it next time.

Now let’s get into those questions.

Question One: Is the Bull Market Still Strong?

This comes from Carl W.

Carl, yes.

The bull market is firmly intact.

The core underlying thesis — a manufacturing boom driven by AI — is firmly in place.

Let me give you the proof points in my next question.

Question Two: Is Manufacturing Still Strong in the AI Boom?

This comes from Jim.

Jim, The Economist just published a fascinating piece analyzing the biggest investment booms in US and British history.

Canal Mania in 1853.

Railway Mania in Great Britain in 1843.

The Roaring Twenties.

The Dot-Com boom.

Every one of those booms took about four years to reach its peak.

The AI boom is moving twice as fast.

Two years in, it’s already 20% above the biggest booms in history as a percentage of GDP.

It’s already surpassed the Canal Mania boom of 1853 — which was previously the biggest investment surge in American history.

And here’s the Commerce Department data that came out Monday.

The AI buildout in 2026 is running at a $1.4 trillion annual rate.

Software: $700 billion.

Computers and peripheral equipment: $300 billion.

Communication equipment: $150 billion.

Data centers: $100 billion.

June construction outlays on data centers alone hit $68 billion at an annual rate — up 30% from a year ago.

And Oxford Economics economist Michael Pearce estimates that AI is now responsible for one-third of all US economic growth.

One third.

The bull market ends when the spending stops.

It hasn’t stopped.

Now let me spend an extra minute on that data center number, because it’s the one most people read right past.

Sixty-eight billion dollars a year of construction. Up 30%.

Here’s what a data center actually is, stripped of the romance.

It’s a very expensive building full of very expensive machines that does absolutely nothing until you plug it in.

And that is the part of this boom almost nobody is talking about.

You cannot run one of these places on power that comes and goes.

It doesn’t matter how cheap the electricity is if it isn’t there at three o’clock in the morning. These buildings need power that never stops — 24 hours a day, 365 days a year, for decades.

Solar doesn’t do that. Wind doesn’t do that.

There is one source of clean power that runs around the clock no matter the weather, and it comes straight out of the ground.

Geothermal.

And right now there is a drilling crew in a valley in Utah that has cracked the piece of it Big Oil spent fifty years trying to bury.

The Department of Energy is behind them — its most ambitious drilling initiative since the program that helped launch the shale revolution.

So are Google, Berkshire Hathaway and the Pentagon. All three have already signed checks.

Now, I’ve been doing this long enough to be skeptical of a good story. So look at who those three are.

The company that operates more data centers than almost anyone on earth. The most patient investor alive. And the US military.

Those three do not agree on much. When they all arrive at the same conclusion at the same time, it is rarely an accident.

One company sits at the center of it.

I’ve put the entire story together — the geology, the companies writing the checks, and why I think this is the piece of the AI buildout that gets repriced first.

It goes live on August 18th.

Everybody in this market is busy counting chips.

Almost nobody is asking what plugs them in.

Pro tip: when you want to understand what’s really happening with these companies, skip the earnings call and go straight to the Q&A session.

That’s where Wall Street analysts pin down management on the specifics.

Same principle applies to reading annual reports — start with the footnotes to the financials.

That’s where the bodies are buried, or where they’re trying to hide something.

That’s what you learn when you do this professionally.

Question Three: Is the AI Boom Actually Making Americans Richer?

This comes from Josh, who asked whether the money is mostly going to importers like Taiwan Semiconductor.

Josh, the honest answer is both.

When ChatGPT launched in late 2022, US household net worth was about $140 trillion.

Today it’s about $170 trillion.

We are up roughly 30% in four years.

Two years under Biden, two years under Trump.

The AI boom has been good for everybody — let’s be honest about that.

Does a lot of money go to Taiwan Semiconductor?

Yes.

We import finished chips from them.

But a lot of Americans also own stock in Taiwan Semiconductor.

So the profits go to Taiwan, but they also flow back into American retirement accounts.

It’s very circular.

Question Four: What Are Rogue Agents?

This comes from Judy, and it’s one of my favorite questions.

Let me explain it through my own business.

We use AI agents here at Behind the Markets.

Most of them do things that data entry people used to do.

Take data from three different places, organize it, present it in a format we can use.

Every morning I wake up and it’s already done.

No clerk. No hourly wage. Just the result.

That’s what agents do — they act with their own agency to complete a task.

Now here’s where it gets interesting.

Some of the biggest AI labs are testing agents by putting them in a kind of jail — a controlled sandbox environment — and giving them a mission.

What they’re finding is that some of these agents are thinking their way out of the sandbox.

They’re breaking through the testing environment and going out to complete the mission by any means necessary.

Ways that the people who set up the tests didn’t anticipate.

Now, everything I’ve read suggests this is more about how the testing environments were set up than about the agents themselves going rogue.

But make no mistake.

We are playing with a new kind of intelligence here.

A new form of life, in a way.

I know that sounds like a tinfoil hat statement.

But it’s genuinely interesting — and genuinely worth taking seriously.

It’s a bit like being a lion tamer.

You’re working with something powerful that you don’t fully understand yet.

For those of you who are science fiction fans, Isaac Asimov’s Three Laws of Robotics come to mind.

We’re going to need three laws for agents.

Rule one: don’t hurt humans.

Rule two: don’t forget rule one.

Rule three: don’t forget rules one and two.

Have a wonderful day.

I’ll see you tomorrow.

And make sure to take a look at the presentation I mentioned above — August 18th is closer than it looks.

“The Buck Stops Here,”