The Nasdaq is at a record, yet 8 of 12 sectors fell. With semiconductors carrying the market alone, how could things play out from here?
2026-W40
How narrow is the market right now?
This is not just a this-week story. Over the past three years, the equal-weight S&P 500 ETF (RSP) has earned 15.9% less than the cap-weighted SPY, including dividends. Only 2.1% of all months since 1929 were narrower than this (The Trading Tools, 2026-10-02). It means the structure in which a few large stocks carry the index while the average stock lags is at its most extreme in nearly 100 years.
There is one more detail the report did not cover. Even the semiconductor ETF (SMH) that pulled the index higher closed at $630.60 on 10/2, 6.14% below its own peak of $671.83. The Nasdaq record was set not by semiconductors as a whole but by a handful of names like Nvidia and Micron. There is a narrower market inside the narrow market.
History points to 'rotation' rather than 'collapse'
The conventional wisdom that a narrow market leads straight to a crash does not fit the data. Here is the performance after the narrowest top 10% of periods (The Trading Tools, 2026-10-02).
| Subsequent period | Cap-weighted median return | Equal-weight outperformance | Equal-weight win rate |
|---|---|---|---|
| 1 year | +16.1% | +0.4%p | 51% |
| 3 years | +3.5% | +4.8%p | 87% |
| 5 years | +3.1% per year | +4.1%p per year | 95% |
Another analysis finds that since 1928, the S&P 500's return in the 12 months after a narrow, leader-driven market was 15.4%, actually higher than the 7.4% after broad rallies (Interactive Advisors, 2020-06). In short, a narrow market says nothing about direction within a year. But it has been a fairly accurate signal that leadership shifts from a few megacaps to the rest over three to five years. That said, the 5-year sample has only five independent periods, so I don't put too much weight on the numbers.
The next two weeks will decide the timing of rotation
Events that will test the semiconductor solo run are clustered together.
| Date | Event | What it tests |
|---|---|---|
| 10/7 | FOMC minutes | Whether long-term yields retest the weekly high of 5.34% |
| 10/8 | Samsung Electronics preliminary Q3 results | A second confirmation of the memory profit cycle |
| 10/13 | JPMorgan results, start of bank season | Whether rate-battered financials rebound on earnings |
| 10/15 | TSMC and ASML Q3 results | Demand across the entire AI chip supply chain |
The TSMC date is based on an external earnings calendar; the company has not yet put it on its official calendar (Akrostec, 2026-10). ASML is also scheduled for the same day.
Three paths
| Scenario | Condition | Semiconductors | Other sectors | Index |
|---|---|---|---|---|
| Solo run extends | Samsung and TSMC beat expectations and the 10-year stays below 5.34% | SMH retests its $671.83 peak (+6.5%) | Still left out | Higher, with breadth narrowing further |
| Healthy rotation | Slowing data pull long-term yields down; solid bank earnings | Pauses below the peak | Financials, healthcare and REITs rebound | Close to flat, but equal-weight leads |
| Joint decline | Hawkish minutes and semiconductors falling on good news like Micron | Breaks below the 10-day MA | No sector to provide support | Falls quickly |
The third is the most dangerous. Right now, if semiconductors take a breather, there is no sector to rise in their place, because this week's bottom three, healthcare, financials and communication services, are all weighed down by rates. Micron's roughly 2% drop the day after strong results is a warning that expectations are substantially priced into semiconductors as well.
So what should you do?
- Semiconductor holders should set warning lines in numbers. A close below SMH's 10-day MA of $607.63 (-3.6% from the current price) is the first warning, and the 50-day MA of $569.85 (-9.6%) is the second. The system's stop-loss line is about $537, -20% from the peak, still far off. A distant stop-loss line is a cushion, but it also means losses grow if you wait until it is hit
- Split new semiconductor buying around TSMC on 10/15. Confirming both Samsung and TSMC would provide grounds for the solo-run-extends scenario
- Index investors should diversify part into equal-weight. Shifting part of your S&P 500 allocation into an equal-weight ETF like RSP prepares you for a 3-5 year rotation, while the performance gap within a year is small. The 1-year win rate of 51% is the cost of that
- Don't sell leaders just because the market is narrow. Historically, returns in the year after a narrow market were not bad. The reason to sell is not breadth but the warning lines above
You can find detailed charts and signals for other tickers on the /signals dashboard.