GDP Breakdown: AI Lost Its Grip on the Economy

The economy just rotated away from AI.

The stock market hasn’t.

During the first quarter of 2026, data pointed to AI representing ~50% of all economic growth, and markets rewarded those stocks accordingly.

But,

Something changed last quarter.

AI’s contribution to the economy fell off a cliff…

NVIDIA, Intel, Broadcom, Micron, and several others have failed to continue their trends since June of this year.

And now we know the reason why,

For the first time since 2023, non-AI growth contributed more to GDP than the entire AI complex:

The headline won’t tell you that.

It actually told you the economy slowed from 2.5% to 2.2% despite this unexpected recovery.

And that creates an interesting disconnect we can massively profit from:

The economy is beginning to rotate, the stock market is still pricing the old story.

You already read about the failed AI economics, particularly Anthropic’s impossible promises leading to over $518 billion in underwater loans.

Instead of kicking AI while it’s weakening, let’s focus on the next set of opportunities.

I want to know where money goes next!

Let’s get into it,

GDP Breakdown: AI lost its grip on the economy.

Let’s compare the first and second quarter prints for a second, and note just how big of a difference there was between:

One contracted slightly, which can be traced to slowing AI expenditures and output, while the consumer suddenly staged an unexpected recovery.

Against inflation and job fears, it was the consumer that really pushed the economy against an AI backdrop.

And there is one specific effect I am interested in, especially as we head into another round of PMI data this week and the next one.

Inventory drawdowns (in gray.)

Because most business operators and managers expected the consumer weakness to continue through the rest of the year, they let inventories get too low.

To a point where this sudden spike in consumer activity ended up drawing them into negative, which creates the sudden spark every investor loves to see.

A spark that boosts production, transportation, and profits higher for everyone in the sector.

This is a trend I had been waiting for, and warned you of in our previous PMI breakdown.

Business complacency around a weak consumer tends to create this inventory imbalance, and when balances get too low… The PMI tends to expand into a new bullish trend.

We’re beginning to see that now.

But,

Who exactly benefits from this expansion, and where should you start looking?

Businesses only invest capital above replacement rates (expansion) for one reason.

They expect predictable demand ahead, and they begin to prepare for it.

Which makes me think this sudden consumer comeback is not a fluke.

More importantly, it’s happening during high inflation and rising bond yields, so imagine just how good it can get when these extremes eventually normalize.

Here were the leaders:

All of which are outside of the realm of AI, and all of which were given to you as an area of interest to find profitable ideas in last month’s PMI breakdown.

Now look at the laggards:

Data centers are in trouble as manufacturing structures take a hit, and don’t even get me started on computer & peripherals.

Memory chips and other analog components for semiconductors were supposed to be in a massive shortage weren’t they?

Then how come they just contracted from ~3.5% last quarter to a negative ~0.5% this quarter?

It’s because most of these companies overshot supply needs, and as data centers begin to slow down and create catastrophic delays, there is simply no need to keep manufacturing as many chips as were previously thought.

That is my entire bear thesis on memory beginning to show itself in the economic data.

Evidence is clearly building against AI, so let’s instead focus on where nobody else is paying attention.

Housing, through a spike in single family activity and consumer strength, is one of my favorite areas to start looking for ideas.

Rest assured, a portfolio is being built around this and will soon be pitched to Offside Premium members.

However,

The coiled spring is not being released at all as you can see above…

We still need to see confirming data out of building permits, housing starts, and real estate specific data like home sales and the like.

The same can be said about transportation and non-AI industrials.

GDP is sure constructive for those sectors, but I would need more specific confirmation through current data (which we may just get this week in PMIs.)

In fact, the only coiled spring that is being released seems to be the consumer:

Notice the contradiction here?

Customers with an already tight budget are choosing to spend in cars of all things, which sounds like things maybe aren’t as bad as they seem on the surface.

I’ve written at length on the effects of the capital cycle, and GDP just confirmed my macro view on this one.

Inventories are rising as a percentage of GDP while capacity utilization remains flat.

Which means inventories are building up because of outpacing demand, not because of an overflow of supply (except for semis/memory.)

Here’s another angle on the capital cycle to confirm what’s coming next:

I want you to take the 2022-Present period of these charts for reference, since that was the last true economic cycle we saw in the United States since COVID.

In 2022:

That led to a broader decline of 20% on the S&P 500 and the last proper bear market we saw since.

Today, we’re seeing the mirror image confirmed through GDP data pointing to a consumer recovery strong enough to drive inventories into the negative.

In other words…

I believe something as boring as trucking, and the consumer will be a more exciting investment compared to AI in the coming years.

Now let’s look at housing:

Except for the 2008 housing bubble, we are at the lowest residential investment to GDP rate since 1992.

Meaning,

There are hundreds of billions of dollars waiting to get back into the housing market as soon as we unlock it, meaning getting mortgages below 6% or so.

Click here to read my view on bonds, it’ll help you time the mortgage recovery.

Bit of a side note…

I am looking to invest in single family homes starting in January 2027, if you live in the United States and know of good deals reply to this email, we can set up a partnership structure for them.

I will keep it simple here.

These are the opportunity sets ranked by highest risk and highest upside:

Now I will visually show you how each name in these baskets will tend to behave in every 1% change in the 10-year bond yield PLUS a rotation outside of the AI concentration scheme:

On the bottom left, you can see most of the housing names with some of the consumers.

Effectively, these are the ones that will outperform everything else if we get a combination of lower 10-year yields AND a rotation out of AI.

Both of which are sort of being supported by this GDP print.

I believe this evidence warrants focusing my attention away from AI ideas and into consumer/housing plays in the coming quarters.

Coming up, we’ll talk more about tomorrow’s PMI and today’s PCE inflation report to piece these ideas together with further data.

And as always…

Fire anyone who told you software was dead, as it was the single largest contributor to GDP outside the consumer!

We will now have to consider the options activity around some of these coiled springs.

I have found an interesting setup in the Builders FirstSource (BLDR) ticker here:

There is a good amount of betting that BLDR will see a massive rally by January 2028, to the tune of ~$192 per share.

Those long calls seem to be funded on a 2:1 ratio by shorting monthly puts for $57 strikes, which is close to where the stock sits today.

Mechanics are simple here…

Dealers will make monthly premiums to pay for their long $192 call strike position into next year, and reduce their timing and movement risk while at it.

It’s the same trade I have on for memory stocks, just betting on the opposite direction.

When you piece it all together,

It’s clear the market has already predicted and digested today’s GDP print, but us mortals are left to figure it out like we did today.

Open this again in 6-9 months and tell me I’m wrong.

Meanwhile, here’s an interview with Bill Ackman bringing up some really interesting points around AI and future disruption uncertainty.

The guy is a billionaire who invests for a living, might be worth listening for a bit:

Until next time,

OFFSIDE RESEARCH

Against the Tape, Ahead of the Curve.