In this market, I increasingly feel that something is off.

Not that it is about to collapse.

But when you look at many macro indicators together, this does not look like an environment that should continue rising mindlessly.

10-year and 30-year U.S. Treasury yields are both elevated.
Global long-term interest rates are rising again.
Oil prices remain high.
U.S. fiscal deficits and debt issuance pressures are still increasing.
Stock valuations are already very expensive.
September is also historically a weak month.

Under normal circumstances, this combination is not friendly to high-valuation tech stocks.

Yet the indexes simply refuse to fall.

Why?

Because the market is no longer trading purely on macro fundamentals, but on expectations of policy support, AI earnings, and the U.S. continuing to maintain asset-price prosperity.

Especially with the U.S. midterm election cycle approaching.

What Trump obviously wants to see is:

A strong stock market.
Stable employment.
Rising retirement accounts.
Continued corporate investment.
U.S. assets continuing to attract global capital.

So this now looks more like a bull market continually extended by policy expectations.

Note that I said “policy expectations.”

That does not mean Trump can directly control the stock market, much less that someone can press a button and push the indexes higher.

What truly supports the market is still corporate earnings, AI capital expenditure, fiscal stimulus, capital flows, and risk appetite.

But that is precisely where the problem lies.

The current bull market is completely different from the most comfortable kind of macro bull market.

A comfortable bull market should look like this:

Inflation declines
↓
U.S. Treasury yields decline
↓
Financing costs decline
↓
Corporate earnings grow
↓
Valuations expand

But now it looks like this:

Long-term Treasury yields are high
+
Oil prices are high
+
Fiscal pressure is heavy
+
Valuations are high
+
AI capital expenditure is enormous

And stocks continue to hit new highs.

This divergence can persist for a long time.

It may even continue squeezing shorts and reach new highs.

But it requires increasingly stronger earnings and narratives to sustain it.

So I will not turn outright bearish simply because the macro environment is unfavorable.

On the contrary, the more a market “should not be rising but keeps rising,” the easier it is to force all the shorts out.

The truly dangerous moment is when the last short seller also starts believing in the bull market.

Going forward, I am watching only three signals:

Whether the 30-year Treasury yield can break above 5.3% or even 5.5%.

Whether oil prices and inflation will spiral out of control again.

Whether demand for AI capital expenditure, HBM, DRAM, and enterprise SSDs will slow down simultaneously for the first time.

Especially the last one.

NVDA is the brain of AI.

MU, SK Hynix, and SNDK are the most direct thermometers of AI demand.

As long as these companies’ orders, prices, and profits continue to grow, this market still has reason to keep extending the story.

But if one day we see:

Treasury yields continue rising
+
AI CapEx is cut
+
HBM/DRAM/SSD demand starts cooling
+
Credit spreads widen

Then it will not be an ordinary pullback.

The stock market did not collapse immediately in 2007 simply because interest rates rose.

The real problem was that after high rates remained in place for long enough, the most vulnerable point finally could not hold up.

Back then, real estate and subprime mortgages cracked first.

This time, I will be watching AI capital expenditure and the credit markets.

So do not rush to call the top.

This market could absolutely continue rising.

It may even produce one final, very beautiful acceleration to maintain bullish expectations.

But I will be riding the gains while keeping an eye on the door.

Because the current macro environment is not telling me “an immediate collapse.”

It is telling me:

The cost of taking this bull market higher is becoming increasingly expensive.