9/22 (Tue) US Market Summary: 'Market Polarization' Behind Nasdaq's Record High | The Tug-of-War Between AI Concentration and 5% Interest Rates
On September 22, the US stock market saw the Nasdaq Composite Index hit a record high for the second consecutive trading day, driven by continued buying of AI-related stocks.
The Nasdaq closed up 0.45% at 27,244.28, the S&P 500 remained largely flat at 7,764.64, and the Dow fell 0.36% to 51,863.69.
While capital inflows into AI-related stocks pushed the indices higher, selling spread to financial stocks and some large-cap technology stocks, making it a day where the direction varied by index. Reuters also reported that while the rise in AI-related stocks and the decline in crude oil prices supported the Nasdaq, the decline in bank stocks and other factors capped the upside for the market as a whole.
However, the most important thing in today’s market may not be the fact that the Nasdaq hit a record high itself.
Behind that, indicators showing market breadth have become quite weak.
The divergence between the ‘index and market internals’ occurring behind the Nasdaq’s record high
Although the Fear & Greed Index rose from 33 the previous day to 35, it remains in the ‘Fear’ territory. It has improved from 27 a week ago, but it was 54 a month ago.
Looking further at the breakdown, both Stock Price Strength and Stock Price Breadth are in ‘Extreme Fear’.
On the 22nd, the NYSE saw 1,335 stocks rise against 1,405 stocks that fell. There were 25 new highs compared to 141 new lows. On the Nasdaq, while advancing stocks outnumbered declining stocks, there were 147 new highs against 178 new lows.
In other words, even though the index is at a record high, the base of stocks supporting that rise is by no means broad.
Meanwhile, the VIX has fallen to 14.21.
Looking only at the volatility market, there is no strong sense of caution, but looking at the advancing/declining stocks and 52-week highs/lows, it appears closer to a state where investors are concentrating capital into specific themes rather than actively buying the entire market.
At the center of that is, once again, AI.
From ‘companies that build AI’ to ‘a world where AI does the work’
In semiconductor-related stocks on the 22nd, Micron Technology rose 5% and Sandisk was up about 7%. Buying also entered NVIDIA, AMD, Intel, and others.
The background to this is the AI agent ‘Muse’ launched by Meta Platforms on September 8.
Muse is not just a generative AI that answers questions; it performs multiple tasks on behalf of the user, such as sending emails, booking travel, and buying and selling products. It is said to have reached 2.8 million downloads in the 12 days since its release, and according to Reuters, it is at a pace exceeding the initial download numbers of ChatGPT during the same period.
This is material that slightly changes the perspective on AI investment.
Until now,
“How many GPUs can we sell?”
“How many data centers can we build?”
I think we have often considered AI demand from the perspective of capital expenditure.
However, if AI agents become fully mainstream, the “inference demand” where AI actually performs massive amounts of processing will increase.
In a world where AI performs a series of actions on behalf of users—searching, comparing, booking, and paying—more computational processing will be required than in a chat that simply answers a single question.
Therefore, it is somewhat rational that expectations for Muse are spreading beyond Meta to CPUs, GPUs, memory, and storage.
On the other hand, there is another aspect to this.
If AI agents begin to step in between users and services, competitive pressure will arise for companies that have previously relied on “customer touchpoints” as a source of revenue.
On the 22nd, Charles Schwab fell over 6%, and there was also selling in Airbnb, Uber, and Lyft. Amazon has stated that it does not allow Muse to access its site.
According to Reuters, investors are beginning to become aware of the possibility that AI agents could replace or intermediate travel bookings, financial services, and shopping.
In other words, this current AI market is
beginning to move from the simple composition of “AI-related stocks will rise” to “a stage of identifying companies that will profit from AI versus companies whose existing customer touchpoints are being threatened”
as well.
The sense of relief in the stock market from crude oil falling below $100
Another factor that supported the stock market was crude oil prices.
WTI is in the $95 range, and Brent has fallen below $100 again.
In addition to the possibility of the Strait of Hormuz reopening being indicated by the Iranian side, the possibility has emerged that Saudi Arabia will resume operation of the East-West Pipeline and restart exports from Yanbu on the Red Sea side. Expectations that supply routes will recover have partially eased the risk premium on crude oil that had been built up due to the situation in the Middle East.
What is important here is that in the current market,crude oil, interest rates, and stock prices are quite strongly linked.
The 200-day correlation of daily changes between the S&P 500 and the US 10-year Treasury yield published by the WSJ has fallen to negative 40.7%, the strongest inverse correlation since 1997.
Crude oil rise
↓
Inflation concerns
↓
Long-term interest rate rise
↓
Pressure on stock valuations
The market is strongly conscious of this route.
Conversely, this time, because crude oil fell, this pressure temporarily weakened, creating an environment where it is easier to buy AI stocks.
However, the US 10-year Treasury yield is 4.958%, and the 30-year yield is 5.296%.
The fact that it is ‘below 5%’ and the idea that ‘interest rates are low’ are not at all the same thing.
It is not just ‘crude oil’ that the Fed has begun to be wary of.
And I believe this is the most important point for looking at the market going forward.
Richmond Fed President Barkin stated on the 22nd that he recognizes the US economy as ‘rather resilient,’ and remarked that inflation risks outweigh employment risks.
What I particularly want to highlight is that they are not viewing inflation merely as a problem of energy prices or tariffs.
In addition to the resilience of consumption, President Barkin pointed out that strength is also visible in defense, manufacturing, and bank lending projects outside of data centers. The Fed raised the policy rate by 25bp to 3.75-4.00% last week, but regarding additional rate hikes, President Barkin said, ‘Whether it is necessary, and how many times, we will see going forward.’
Boston Fed President Collins also supported this rate hike and expressed the view that the risk of inflation continuing to exceed 2% is increasing.
In FedWatch, the probability of pricing in an additional rate hike of 25bp or more at the October meeting has risen to approximately 53%.
In other words,
if crude oil prices fall, it does not mean the Fed’s problem is solved
.
If the inflation seen so far were an ‘energy shock due to the Middle East situation,’ the Fed would have room to temporarily wait and see.
However, if demand itself is strong and companies pass energy costs and tariffs on to prices, and that spreads to service and product prices, the story changes.
The ‘Second Strike of crude oil’ pointed out by the WSJ is also here.
Rising gasoline prices first push up headline inflation. If it then spreads to transportation costs, manufacturing costs, and service prices, it will also affect core inflation.
Unlike the time of Russia’s invasion of Ukraine in 2022, this rise in crude oil prices has remained significantly above pre-conflict levels for over 200 days, and one must be cautious about the possibility that companies will begin to perceive this not as a ‘temporary cost increase’ but as a ‘permanent cost.’
Even with stock prices at record highs, valuations are actually declining.
On the other hand, it is also slightly incorrect to simply view the current stock market as ‘overvalued.’
According to LSEG, the S&P 500’s forward P/E has fallen to just under 19x, the lowest level since 2023. The background to this is not only the adjustment in stock prices but also the upward revision of earnings forecasts, centered on AI-related companies.
While the S&P 500 is just a short distance from its all-time high, its valuation based on expected earnings is declining.
This is an interesting aspect of the current market.
If stock prices rise while EPS forecasts grow at an even faster pace, the P/E ratio will decline.
In that sense, what will be important going forward is not simply whether the index is high or low, but whether corporate earnings can keep up with current expectations.
If AI investment leads to actual sales and profits, it can support current stock prices.
Conversely, if AI capital expenditure and the monetization of AI agents do not progress as expected, current earnings forecasts themselves may be revised.
A market where ‘record highs’ and ‘fear’ coexist
The Nasdaq has hit a new record high.
However, the Fear & Greed Index is at 35.
Market breadth is weak, long-term interest rates are near 5%, and the Fed has not ruled out additional rate hikes.
Even so, capital is returning to AI and semiconductors.
At first glance, this seems contradictory, but it may well represent the current market accurately.
It is not that investors have started taking on risk across the board, but rather that
they are concentrating the room created by falling oil prices and slightly stabilized long-term interest rates into AI-related stocks with high earnings growth potential.
Viewed in this light, the Nasdaq’s record high and the ‘fear’ in the Fear & Greed Index are not necessarily contradictory.
A US-China summit is scheduled for this week, and discussions surrounding trade, AI, and geopolitics will become new market factors. According to Reuters, topics such as AI and the extension of a trade truce are being watched, but the specific details of any agreement remain unclear.
When looking at the current market, it is not just about whether the Nasdaq can continue to hit record highs, but also
whether oil can stay below $100, whether the US 10-year Treasury yield will exceed 5% again, and whether the rally will spread beyond AI stocks.
I would like to check these three points together.
Looking at the index alone, it is strong.
However, beneath the surface, many stocks are still far from their record highs.
Rather than saying the “stock market is strong,” it might be easier to understand the current market by viewing it as a market where capital is concentrated in strong areas.