3 Market Shifts You Should be Watching Now

Something big is happening underneath the AI trade.

Chinese models are quietly taking over the entire market share of token usage, and up to 50% of all large U.S. technology companies have chosen Chinese models rather than American ones.

Even 80% of small businesses and startups, according to Morgan Stanley, have decided to go this route instead.

And the reason is not difficult to understand…

Chinese models are roughly 80-85% cheaper than their American competitors, while the capability gap has narrowed down to a point where tasks completed are nearly identical.

I’ve been testing some of these models myself, migrating my context data from Claude and ChatGPT into Chinese AI instead.

The experience? Surprisingly similar.

The bill is not.

But here’s where the story gets interesting.

Despite all of the market share and token gains, most Chinese tech stocks are down nearly 30% this year and still trade at a massive discount to U.S. tech stocks.

That is market shift number one.

The second shift is happening inside the financing behind the AI boom.

Risks are no longer found in chip availability, as you know from my Micron earnings analysis that the market will be flooded with chips in the coming months.

Debt and credit are becoming the biggest risk now.

AI-debt issuance collapsed from $113 billion in June down to $23 billion in September as the cost and risk of financing has spiked accordingly.

Some of the largest infrastructure and compute commitments are at risk of default.

And that risk goes up every time token prices go down, leading to OpenAI and Anthropic to miss their revenue targets.

That changes how you should think about the AI trade as a whole.

Then there is energy.

Crude shipment volumes are back to normal, but refinery volumes aren’t.

As you can see, oil prices are beginning to normalize a bit while diesel remains near all-time highs…

That’s a good thing!

It means it’s just a matter of time before refinery outputs pick back up, and available supply catches up with prices.

That helps our long/short equity trade that pays on falling diesel prices.

Today I’m breaking down:

More importantly,

You will be routed back to the theses and ideas behind these thematic investments.

All of which are being proven right so far, and profits are soon to follow.

Let’s get into it,

3 market shifts you should be watching now.

No industry is safe from the laws of the capital cycle.

When a select group of companies begin to make outsized profit margins offering one product or service, you can be sure that lots of new participants will enter the space and compete.

That is exactly what we’re seeing in the AI sector.

In 2022 it was just ChatGPT laying the foundations of what we see today, then Claude joined and all the sudden we had access to over a dozen different AI models and services.

Today, the market has fragmented into:

In other words,

China has passed the U.S. in global AI market share, token usage, and customer migration outside of the leaders like Anthropic and OpenAI.

Now here’s where everything you’ve seen recently begins to make sense…

Elon Musk, Sam Altman, and Dario Amodei all called for the global regulation of AI.

They even cited scenarios where we all end up without jobs and dead if we don’t manage to control the growth and development of AI.

Days later, these AI giants went onto raise billions of dollars in new capital.

It’s all a big pile of lies my friend.

The reality is that these AI companies are losing so much market share to China that the only way to stop the bleeding is through regulation, which is also the only way industries can escape the capital cycle.

Tobacco, big oil, big pharma… They are all heavily regulated because of this same reason.

There’s one huge headwind to regulation though, and the reason why President Trump chose not to pursue government involvement in the matter.

This is not an opinion or bias, it’s the actual data made public by Anthropic itself.

If the government gets involved at this stage, they are exposed to taking the blame for the industry’s financial collapse, and that would not be advantageous whatsoever.

Because with blame comes financial responsibility, a bailout if you will.

Some more data on the topic from Morgan Stanley:

This is where the opportunity is born.

As bearish as I may sound at times, I am completely convinced that AI will be as big as the internet, but that doesn’t mean all companies win.

For the first time this century…

China may be the one to win the big technological revolution, and you can ride this decade-long wave at a discount:

Chinese tech firms trade at a ~50% forward P/E discount relative to U.S. tech firms.

Most of which are valued before EPS adjustments, as roughly $0.60 for every $1.00 in EPS is now made up of accruals (non-cash items) that could easily be wiped away overnight.

I will close the topic with the promise of an in-depth study on the Chinese market, economy, and investment deals to bring to you.

Let’s now take a look at how all this China rotation is affecting AI credit markets in the U.S.

2026 had so far been a booming market for debt issuance related to AI capex and data center buildouts, to a point where it hit a record 12% of all investment-grade bond issuance.

However,

That all changed starting in July, where AI-linked debt issuance fell from $113 billion to $46 billion.

Then again in September, where it halved to $23 billion.

Timing coincides with all the KPMG and Morgan Stanley studies showing China began taking over the AI complex at a global scale, directly threatening the economics of AI within the U.S.

A few things also led to this sharp slowdown in debt issuance:

If you’re an investor betting on AI’s momentum, this means your expected payoff got pushed back three years or more.

In other words,

An expected 50% return is now closer to 15%, and you still are on the hook for paying interest each month these projects are delayed.

This risk has become so large that the rest of the debt market is falling into contagion risk.

High-yield credit spreads are spiking once again, which is a direct gauge as to how risky investors think the credit profile of the U.S. economy and the S&P 500 is right now.

What started as an AI-only risk has grown to become a massive threat to the financial system.

Which is also why markets stopped treating AI companies the same way:

Neoclouds, Chipmakers, and asset managers now trade in tandem with high-yield credit.

Meaning,

These stocks are no longer driven by how many chips or compute contracts they can sell, but rather on how much AI financing they can issue and attain.

The reason is part of those accrual items I mentioned above.

When balance sheets and earnings become a portfolio of circular financing from NVIDIA and GPU-backed loans, your company essentially becomes a credit play rather than a native AI play.

I broke down every single figure, risk, and exposure to these shady accounting standards a few months ago.

Click here to read my forensic analysis on these companies.

In the Hormuz region, crude oil exports are back to the same pre-war levels now, or roughly 2% within them.

What hasn’t caught up is the volume of diesel and gasoil exports, which are still just 26% of what they used to be before the war broke out.

The reason is simple,

These refineries were put offline in the Middle East and in the U.S. alike for quite some time, affecting diesel supply and pricing across the world.

But,

Now that oil is beginning to flow freely, it is only a matter of time before the improved economics work their way toward diesel, as available oil leaves enough excess supply to put toward refining activities.

In my view, this is going to help bring diesel prices down in the coming months, which is exactly what I was betting on before the November midterm elections in the U.S.

Click here to see how I’m choosing to make money on this idea.

What was driven by the diesel timing crisis is now becoming the initial normalization stage I just broke down above.

I’m confident in taking this view as the WTI cost wedge remains relatively flat along with gasoline, meaning the only tail risk that is present can be centered down on diesel refinery timing.

And the market is already confirming this take:

U.S. refineries have already taken most of the gains since the war broke out, with Asian players now running into flattish channels and even some declines.

What this means is there is an outsized excess return to be corrected in U.S. refineries, which I believe could be about to unfold given how the market is now selling diesel futures to bring these quotes lower.

My selection on how to play this has already been made available to Premium members, and I will continue to update them through the month on how the China, AI, and fuel theme positions are managed.

A lot of coverage has been given here today, so I will point you to some resources you can check out for this opportunity set.

For the AI complex turning over, I am choosing memory as my main bearish bet.

As you know from my Micron positions last quarter, timing this theme is nearly impossible…

So I will continue to suggest a long/short equity or options strategy to ride this peak and turn cycle.

For the rest, refer to the following:

Now for domestic AI and memory options:

A similar spread to the one I already have in the portfolio continues to be held by dealers.

Short the October-December puts at $980 strike, use the credits to buy twice as many long October-December puts at $600 strike.

I have personally had this trade on since August, and will be rolling my short legs this week on options expiry, potentially picking up a new round of long puts dated in 2027 to expand on this theme.

You can piggy back on this idea, or you can cut the risk of mismanaging the spread by joining Offside Premium.

If you’re already part of the community, then be on the lookout this week as I execute on this trade.

Meanwhile, here’s the latest from Jordi Visser (the AI retail voice) to update you on what the bulls are leaning on as far as their narrative.

Take it with a grain of salt, question it, and see for yourself just how strong/weak the trade is right now:

Until next time,

OFFSIDE RESEARCH

Against the Tape, Ahead of the Curve.