The Bubble Has Popped

The price action over the past 2 months has been extraordinary.  You rarely see such a divergence between SPX and NDX.  The NDX is trading almost 2 times the beta of the SPX.  What makes that even more unusual is that the NDX components, the Mag7 as well as MU and AVGO, make up 9 of the top 10 stocks by market cap in the SPX.  The top of the SPX board is dominated by NDX names, yet you still see a huge difference in price action between the 2 indices.  The last time you saw this was in the spring of 2000, when the dotcom bubble popped.  It looks like the AI bubble popped in June 2026.  

There has been quite a bit of volatility both inter-day and intraday in the SPX, even though you are close to all time highs.  Usually the VIX is mostly between 13 and 16 when near all time highs.  This time, its hovering mostly between 16 and 19.  Volatility increases around market tops, as you get more selling to match the buying, causing sudden intraday drops that you rarely see in the middle of a bull market.  

This market is giving you a lot of time to buy around lows since the June top.  In a bull market, they don't give you much time to buy near the lows.  In a non bull market (both bear and sideways markets), they give you a lot of time to buy near the lows.  Some examples below.

2000.  Before the start of the bear market in September 2000, the SPX gave traders plenty of opportunity to buy local lows well below all time highs.

2015-2016.  Before the waterfall declines in August 2015 and January 2016, SPX had several pullbacks of 3-5% while being range bound.  

2021-2022.  This top was a bit trickier, as the topping process was much briefer than in 2000 and 2015.  Only real warning signs of impending top was the November/December weakness, but the false breakout to 4800 probably would have hit a lot of stop orders for premature short positions.  The waterfall decline in January 2022 gave you many opportunities to buy near the lows, signifying significant weakness.  The late March top after the counter trend rally was a great shorting opportunity.  

2026:  The SPX appears to have hit a ceiling at SPX 7600, now going on 2 months without making a new all time high.  More importantly, spending a lot of time near local lows, with no V bottoms during those 2 months.  Super high dispersion and a very split market is similar to what was seen in 2000.  NDX relative weakness after the early June top is another tell, similar to the post March 2000 top.  

So if the AI bubble has popped, why not be short?  Before this week, there was TACO risk on the short side, as Brent crude went above 100 and there were war fears percolating. There was still potential for a reflation of the bubble to early June levels, after another TACO.  I was expecting a relief rally after the TACO, big tech earnings, and the FOMC meeting.  You had the reflexive rally after the TACO over the weekend, and it lasted just a few hours during the overnight session.  The relief rally finally happened after all the events had passed on Thursday, but only after getting pounded down heavily ahead of them.  These gradual, extended selloffs and sharp intraday selling of gap ups are hallmarks of weak markets.

After seeing how sharply the AI hardware and semiconductor names went down for the first 3 days this week on no news (even some good TACO news), while the rest of the US market was barely negative, it showed how saturated the AI positioning has gotten.  Some may say it was Leopold Aschenbrenner's large AI fund's blow up that caused the cascade of selling, but his fund wasn't large enough to cause such a huge move.  And if that is all it takes to crush the AI names, then that is a show of extreme weakness itself.   

Looking at the put/call ratios, you didn't get the panicky readings that you had in June at those local bottoms.  It is a bit surprising how unaffected the SPX is considering how much the high beta tech leaders are faltering.  It is deja vu, but 26 years later.  

Retail heavily selling single stocks this week, net daily flow hit -243M on 07/28.  Retail has gone from buying single stocks (speculative) to buying equity ETFs (less speculative).  Still speculation, but going from high beta tech speculation to just beta equity index or somewhat high beta tech fund ETF speculation.  They are still heavily invested.  The next stage of the selling will happen when retail goes from buying equity ETFs to selling them to go to bonds or cash.  That's when we get the waterfall.

 

This week, you saw some retail traders throw in the towel, as the DIX went down towards 40, a level that has marked local bottoms in the past.   

 

Bigger picture, heavy foreign investments into US equities is a warning sign of a late bull market.  Peaks in 2000, 2007, 2021.

Insider buying is extremely low.  

Bonds continue to trade weak.  Hitting new highs in 10 year yields on Friday.  Yet we still have investors acting like ostriches burying their heads in the sand.  Expecting the Fed to stay on pause, or even expecting the next move to be a rate cut!  Recency bias is prevalent, as they are expecting another no-reason rate cut, just because Trump wants it.  They could be in for a shock when the rate hike happens later in the year, as is being priced in STIRs.  

COT data as of 07/28 showed asset managers buying the dip (unusual) in SPX and NDX futures, while dealers added to net shorts in both.  Small speculators continue to maintain large net long positions in both.  

There was more evidence this week that the final top of this bull market for NDX was reached on June 3.  The relative weakness of NDX vs. SPX has extended out towards 2 months.  It can't be shrugged off as mere rotation.  There has been no bad news in the earnings outlook for AI names to justify the weakness.  Some may view that as a irrational selling.  I view it as extreme weakness on no news, which is a tell.  Funds have overallocated to the AI trade, and the incremental flows are not there to support these bubbled up prices.  

Mostly in cash, and waiting to see if we get a short term oversold bounce that takes SPX to between 7550 to 7600.  That is a prime area to enter short positions in the index.  Individual tech names popular among retail are also good short candidates on a bounce.  This is not like 2025, when you had a big flush out on Liberation Day and it took several months to get to extended bullish positioning, and high beta names/retail favorites outperformed throughout.  In 2026, you had a mini flush out due to the Iran War, but it only took 2 months to get to extended bullish positioning, and high beta names/retail favorites are lagging for the past 2 months.  Its a much more bearish picture in 2026, and it is something that investors are only just beginning to realize.  The real selloff happens when the point of recognition hits later in the year.