PMI: A Stock-Picking Cheat Sheet

Global M&A activity, combined with what you’re about to see for August’s PMI readings, can tell you a whole lot about where capital is already concentrated.

And, where it may move next.

Two things to keep in mind as you digest this:

AI-related technology dominated M&A activity, with most transactions happening in the $10 billion-plus range.

On the other end, consumer and retail saw the least amount of deals, each happening in the $5-$10 billion range.

Why is this important to know?

M&A tends to accelerate when valuations are rich, and slows down when prices compress.

PMI can show us where these two extremes begin to converge and rotate.

When that rotation begins, that is where stock ideas get interesting.

Especially if one group is crowded, expensive, and heavily financed…

While another is quietly reporting better demand, rising activity, and overly pessimistic earnings expectations (making them easier to beat.)

That is what I want to find today.

If you haven’t, click here to read how this capital cycle affects memory companies specifically.

Let’s get into it,

PMI: A Stock Picking Cheat Sheet.

One thing you’ll need to know about the two PMIs:

The AI buildout has disturbed the balance of these two reports, as an eight-consecutive month expansion in manufacturing can be accredited to AI-related materials.

More importantly,

This level of activity translates to 50% of 1H26 GDP growth coming from AI infrastructure investments and AI-related exports.

Which means:

As the S&P 500 is a reflection of GDP, you’ll now understand why the market has become overly concentrated on AI names, and why correlations have spilled over to several industries outside of technology.

Why am I telling you this?

Because manufacturing contracted on almost every important metric, while services pushed ahead.

And because most of manufacturing activity is made up of AI-related projects…

Could that mean the rotation is soon to happen?

The top industries behind manufacturing’s shift are:

These have now created a few themes to consider as we begin generating ideas off this report.

Such as:

Steel and aluminum are responsible for that primary metals expansion, which itself is driven by shifts around tariffs.

In fact, this is a supply driven expansion rather than a demand spike!

Which is why we played and exited our Cleveland-Cliffs (CLF) earnings swing in less than two weeks last quarter.

Still,

The combination of Section 232 tariffs on metals and automobiles, falling used car prices, and rising interest costs to finance new cars…

All lead to a reiteration of a potential boom cycle in the automotive industry for the U.S.

As you’ll see in my OPENLANE (OPLN) report, this is something the market is already positioning for through forward P/E premiums.

As far as the other industries go, specifically centered around AI…

We can see a sharp decline in new orders from the electrical equipment space, which points to a potential decline in demand intensity for these input components.

Note: This doesn’t mean demand is gone, it’s just less intense than in previous months.

Computer products share a similar trend, cooler demand than in May’s AI peak, but not weak enough to warrant a shift in the sector just yet.

Here’s another reason to remain flexible around AI right now:

Even though computer and electronic materials have been in shortages for 8-18 months, demand has snapped its heating trend.

In my opinion this was bound to happen.

GPU and memory makers continue to raise prices, like NVIDIA’s 15% hikes last month.

At some point, extending lead times and more expensive products will make the end user think twice about expending these costs, which is exactly why I think new orders have slowed down for AI-related areas.

Services confirms two themes hinted inside the manufacturing report:

Wood products in manufacturing have seen the worst quarter on record, and construction isn’t doing great either.

Which makes sense as mortgage rates continue to go up, pricing out most would-be homebuyers into the sidelines as a result.

Therefore, this real estate expansion in services is more due to leasing (if people can’t buy, they rent) as well as the growing square footage being leased for data centers coming online.

Recall electronic and metal components related to the AI buildout are in a year-long shortage, and the services real estate activity is further confirmation of that.

One thing we can take from this chart as pure signal:

Which translates to a ton of data center (and compute) supply coming online in a single month.

Perhaps this is one of the reasons why Broadcom, Marvell, Oracle, CoreWeave are all in a sideways market.

As new supply comes online, pricing power is at risk of deteriorating, especially now that demand for electronic and computer components seems to be softening.

Wholesale and transportation are an interesting story here too…

Tariffs, which are pushing steel/aluminum prices higher with a small yet present tailwind from automotive demand, have created a vacuum in these two industries.

Trucking and logistics names are in the eye of a rising activity storm.

Manufacturing confirmed this through transportation equipment.

If I had to say anything about the PMIs, it is that they strongly reiterate my view on two of my recent picks:

They also expanded my idea pipeline into other transportation areas like trucking, coupled with some primary metal makers in the steel and fabricated metal industries.

But above all…

A weak consumer is still choosing to spend in both convenience and experiences (as seen in accommodation activity.

Convenience and experiences that are part of my thesis behind:

Which is why I guess they have done well so far since our buying price, and why they continue to hold even through market chops.

There’s a classic options spread going on for KnightSwift Transportation (KNX) as follows:

Participants have accumulated a very particular bet for December 2026:

Not a typo, the same intensity is going on a 1:1 ratio for the same strike and date.

This closely resembles a one-way directional bet on KNX stock rallying past $105 by December.

If things continue to develop as they are in the PMI data, I believe this is one such trade that pays off very well.

You short the $105 puts to finance a lot (if not most) of the long $105 calls.

By the time the stock hits $105 or above, your shorts will be near zero while your longs will be in deep profits.

Not something I am particularly putting on, but keeping an eye on as confirmation that the data is pointing me in the right direction.

The actual trades in trucking are being sent to Offside Premium members once the structure is done with.

Meanwhile, here is the latest from Jordi Visser as the retail authority voice on AI.

You will find a lot of his data and views seem to be conveniently focused on all the good, and none of the obvious negatives:

Until next time,

OFFSIDE RESEARCH

Against the Tape, Ahead of the Curve.