Smart Money Isn't Hedging, They're Selling

Last week, a BofA research paper landed on my desk.

I wasn’t expecting much, given these are usually noisy and generic.

After digging into it, building my own model, and comparing it against every major market peak of the last three decades…

I realized something uncomfortable.

The market is sending one of the strongest bullish signals I’ve seen this year.

Ironically, it’s also building the conditions that preceded every major correction in modern history.

Both can be true, the question is for how long.

The answer lies in an institutional concept most investors never get to see:

Dispersion.

Today we’ll cover:

This isn’t a buy or sell signal for stocks in itself.

Dispersion tells you how much energy is building in a sentiment coiled spring, and how aggressively it can pop.

Timing that pop isn’t your job.

Your job is to be prepared, and line up your portfolio ideas according to the probability of these events happening.

Let’s get into dispersion,

The reason why smart money isn’t hedging, they’re selling.

Every investor gets introduced to the fear index (the VIX) at some point in their career.

What they are never shown is the VIX’s hyperactive cousin, the VIXEQ.

While the VIX covers the S&P 500 implied volatility, or the consensus everyday vote on where people think volatility will be on an annualized basis.

The VIXEQ covers that same consensus, only at the individual stock level.

Meaning,

If the S&P 500 is truly diversified, then the single-stock volatility should be very close to the index-level volatility.

We can appreciate this in the chart above before the divergence of 2026.

That divergence is the reason for my research, and why we’re here today.

Dispersion is defined as a two-way bet where participants will buy VIXEQ exposure while selling short VIX exposure.

In a way, they’re betting on concentration and excess leverage delivering returns at the individual stock level, while a lack of diversification in the S&P keeps the VIX low enough to make money on the shorts.

Short the VIX → Receive funds on margin → Buy VIXEQ → Profit from AI concentration

That’s the entire dispersion trade today.

Now let’s get into the second aspect that allows a dispersion trade to take place:

When the VIX is low, correlations tend to thin out too.

That’s breeding ground for dispersion to build, as low volatility allows for maximum risk-taking in whatever the hot bet is for the day.

We saw extremely low correlations in 1998, leading to a dispersion trade to build around the internet names.

Then in 2007 as dispersion built around homebuilders and mortgage originators.

Again in 2014 energy trade dispersion, leading to the 2018 “Volmaggeddon” episode and a subsequent spike in COVID.

2022 counts as correlation swings and VIX seasonality (this time without dispersion.)

Right now in 2026, record-low volatility readings have created the AI dispersion as participants see a growing need to chase returns in whatever is moving.

To recap:

The 3-month rolling dispersion index (DSPX) shown in Thinkorswim is distorted by recent events.

After Citadel bailed out Situational Awareness’ positions in most of the concentrated AI names, their volatility compressed significantly.

Which is why the DSPX index declined so aggressively.

However,

Now that Leopold Aschenbrenner got another $400 million injection, running up leverage like never before and buying back everything he had to sell…

The 1-day realized dispersion index is back to a new all-time high (past the dot com bubble.)

Mechanically, this is bullish for most of the AI trade.

One reason why I decided to buy a select few, while shorting another for risk-management.

Coincidentally, whatever is bullish for the AI trade is also bullish for the S&P 500 index.

Which is why, despite all of the deteriorating data we’ve been covering, I just cannot turn bearish on the index just yet.

The reason is simple,

As long as dispersion exists (and keeps widening), more money will pour into the trade and keep it racing higher.

But,

Not being bearish doesn’t mean I am blindly bullish on this fact alone.

Like I said before, dispersion doesn’t mean buy/sell, it merely measures how coiled the sentiment spring is right now.

The read?

We are at unprecedented levels of concentration, VIX short interest, and bullish sentiment.

Now we get a bit more technical.

Whenever the dispersion trade deviates toward a 15% - 20% Z-score deviation, a broader S&P 500 correction tends to happen.

Again:

In other words,

This is the most bullish (yet most bearish) market we’ve ever had in the past 30 years!

Statistically though,

Each deviation beyond 15% registers an S&P pullback of roughly ~20% off the highs.

Essentially a bear market once dispersion gets too wide.

At a current 30% deviation, I would suspect the pullback from highs should be much more aggressive.

Because any mean-reversion from dispersion would automatically force VIX short sellers to buy the index in bulk, mechanically wiping out hundreds of billions in AI-related longs.

Remember dispersion is merely a gauging tool (not a timing one.)

Yet,

It seems we have reached such extreme levels, that some market participants have decided to proactively reduce exposure ahead of whatever event this may bring.

In my weekly plan, I specifically showed you how the market seemed to rally without much volume nor factor support.

In other words,

The market climbed without AI names nor futures buyers to back the move.

In fact, the commitment of traders report showed both leveraged and managed money sold exposure in the index together.

Usually, managed money (pensions, mutual funds) buys while leveraged money (hedge funds, prime brokers) sell to hedge or raise margin.

When they both sell together, it means the market has become too risky, even for them.

I believe this goes beyond the dispersion trade and into the overall capital cycle for semiconductors, infrastructure, and memory.

Stay tuned as due diligence is complete.

Shamelessly copying the take from last week, as it seems to be the best way to gauge whether dispersion breaks or supports the market ⬇

If the computer and electrical equipment theme continues to heat up in the second half of the year…

I would like to see the SMH ETF break away from the previous overhead resistance made, which represented a 50% retracement from its last leg lower.

So far, we remain in a lower-low lower-high pattern, and that’s all you need to know.

Breaking this ~$590 resistance could reignite the path toward a new all-time high, especially as the GDP and PMI data continue to favor an EPS upgrade for these companies.

However,

As I mentioned in this piece, you need to become more selective in your screening.

Were the ETF to fail despite continued PMI expansion, I believe stock selection alpha will be the only thing to carry you into the green from here.

This trade is the easy take as long as it breaks.

If it doesn’t…

You know where to find alpha, as I delivered some already earlier today:

Click here to see our latest trade alert for the AI rebound.

Now here’s an addition for this week, backing the dispersion study we just did:

Volatility remains very low for the SMH.

Its rate of change has even swung into a negative deviation band, meaning we should probably see a rebound soon.

My Commitment of Traders take usually presents itself as a slow grind before a big move (AKA late 2025.)

So perhaps we will see continued buying and a breakout on SMH to reach the upper bands of this chart again.

Either way, my AI trade is protected.

Meanwhile, here’s the latest from Jordi Visser, the authoritative voice in the retail chase for the AI trade.

Perhaps we align on some things here short-term, though long-term I am convinced a lot of people will complain about following his advice:

Until next time,

OFFSIDE RESEARCH

Against the Tape, Ahead of the Curve.