Retail Trading Just Broke Every Volume Record — Here's What the Data Shows

And it does concentrate. Retail traded roughly $1.9 billion a day in semiconductor options premium in June — six times the historical average, with about three-quarters of that activity sitting in calls, according to Citadel. Nasdaq rallies in May came paired with rising implied volatility around 70% of the time, the highest monthly reading since 2005 and more than triple the long-run average — the opposite of how volatility normally behaves in a calm uptrend, and a sign that traders were paying up for continued upside rather than treating the rally as safe.

The concentration shows up in the index math too. Semiconductor companies made up about 18% of the S&P 500 as of July 2026, up from 12% a year earlier, 5% five years ago, and just 3% a decade ago, per Citadel's July 13 note. Three-month implied volatility across the ten largest chip names has climbed from 29% in 2016 to nearly 73% now — the options market has effectively re-rated the entire sector as structurally more volatile, not just temporarily hot.

Anyone who traded through the 2021 meme-stock period will recognize the volume numbers but not the pattern underneath them, and that distinction matters more than the headline stats do.

2021's retail surge concentrated in idiosyncratic, high-short-interest small caps — the trade was finding the next overlooked name before the crowd did. Today's retail flow, by Citadel's own framing, increasingly mirrors institutional positioning: the same mega-cap and semiconductor names that dominate index returns are also where retail options premium is piling up. That's a fundamentally different behavioral pattern. It's less "find the hidden gem" and more "pile into the trade everyone already agrees on" — which changes the risk profile entirely. A crowded idiosyncratic trade can blow up one stock. A crowded consensus trade concentrated in the names that already drive the index can amplify moves across the entire market at once.

It's part of a broader pattern, too — the same retail appetite for fast, concentrated bets shows up in prediction markets and event contracts, which posted their own record volume this same stretch. Whether it's a same-day options contract or a binary event contract, the through-line is a retail trader increasingly drawn to short-dated, high-conviction bets rather than longer-horizon positioning.

One statistic captures just how unusual the current setup is: three-month implied correlations between individual stocks fell to their lowest level in 15 years, per the June 30 report. Read literally, that says stocks are moving less like a herd and more independently of each other — genuinely good news for anyone whose edge is picking individual names rather than betting on index direction. Read alongside the semiconductor concentration data, though, it also means a huge and growing share of retail dollars is betting heavily on outcomes for a fairly narrow list of related companies, even while the broader market fragments into more independent stories.

Citadel's freshest note, published July 13, adds real-time texture to the trend rather than just more of the same records. As of that date, the firm's retail platform had not logged a single net-sell day for the entire month of July. Daily net buying was running at roughly 3.2 times the historical monthly average, putting July 2026 on pace as the second-strongest month for retail net buying since January 2020 — and the strongest July the firm has on record. The final week of June into the first week of July produced the two strongest weeks of momentum-factor retail buying the platform has ever measured, with July 1 alone seeing net buying run nearly 12 times the trailing one-year average.

There's a genuine crack worth watching, though, and it's the first one in months. Retail sold semiconductor and hardware names on the SOX's two most recent down days — a real break from the buy-every-dip pattern that had defined the entire cycle up to that point. Both selloffs were followed by sharp rebounds, so it hasn't cost anyone money yet. But it's the first sign that the semiconductor trade, the single most crowded corner of this entire retail surge, might have a limit to how automatically it gets bought on weakness.

Some of the mechanical pressure that built up earlier in the summer is also easing, which matters for anyone worried this rally is purely a leverage story. Leveraged ETF assets had swelled to a record $218 billion by late June, but had already contracted to roughly $198 billion by mid-July — a pullback of about 10%, concentrated mostly in semiconductor-linked products, which fell closer to 20%. Smaller leveraged ETF assets mean less forced, mechanical rebalancing at the close each day, which typically means fewer artificial end-of-day price swings. Financing costs have eased too: one-month equity financing spreads, which had spiked to 138 basis points over SOFR, had come back down to around 60 basis points by mid-July, making it cheaper to hold long positions on margin.

The next two weeks carry real event risk on top of all this. Roughly 36% of S&P 500 weight and 33% of Nasdaq 100 weight are scheduled to report earnings between July 27 and July 31, including four of the Mag 7 — Citadel's own framing calls it the heaviest single week of the entire earnings season. Consensus sits at 22.4% year-over-year EPS growth for the quarter, a number that would rank among the strongest readings outside a recession recovery if it holds. Semiconductor companies, unlike most sectors, report throughout the calendar rather than in one concentrated week, which means the elevated implied volatility in that corner of the market isn't going away once late July passes — it stays elevated into August.

Three practical things fall out of all this if you're actually trading intraday rather than just reading about it.

First, liquidity and spread behavior in mega-cap and semiconductor names is now shaped as much by retail options flow and market-maker hedging as it is by traditional institutional order flow. If your setups depend on reading volume and price action in names like the largest chipmakers, you're increasingly reading a market where a meaningful slice of the volume is same-day options hedging rather than directional conviction. Our relative volume guide is worth revisiting with that context in mind — a volume spike in a heavily 0DTE-traded name doesn't necessarily mean what it used to.

Second, the correlation breakdown is a real opportunity for anyone whose edge is stock selection rather than macro direction. When individual stocks move more independently of each other, strategies built around finding the right name — not just riding the tape — get more room to work. Our guide to identifying the market regime covers how to read whether you're in a stock-picker's market or a macro-driven one; right now, by Citadel's own correlation data, it's clearly the former.

Third, the leadership broadening documented in the July 13 note is worth tracking on its own. Since early June, a majority of S&P 500 constituents have finished higher on the index's down days noticeably more often than the trailing-year norm — a sign that money rotating into previously lagging sectors like Communication Services and Financials is cushioning broad selloffs rather than every stock falling together. If you scan for sector rotation setups, this is exactly the environment our sector rotation strategy guide is built for, and a scanner that can flag relative strength shifts across sectors in real time — something a platform like Trade Ideas is built to do — earns its keep in a market moving like this one.

None of this is a reason to chase the trend blindly, and it's worth being direct about why.

A market this concentrated in a handful of retail-favorite names carries a specific kind of fragility. When a huge share of trading activity is chasing the same semiconductor names, wrapped in the same short-dated option structures, a single disappointing earnings print or guidance cut in one of those names can trigger outsized, correlated selling across the whole group — precisely because so many traders are positioned the same way. The two recent SOX down days where retail actually sold, rather than bought the dip, are a small preview of what that looks like when the pattern breaks.

0DTE concentration carries its own behavioral risk independent of the market structure story. A contract that lives and dies in a single session rewards constant checking and fast reflexes, and it punishes hesitation just as fast. If you already recognize overtrading tendencies in your own history, a market environment saturated with same-day options activity is not the place to test new limits — the instrument is specifically built to compound impulsive decisions quickly, in both directions.

And a general caution applies to leaning too hard on any single data provider's framing, however good the access. Citadel's numbers describe what already happened through mid-July. They don't guarantee the pattern continues through the earnings gauntlet at the end of the month, and a firm with a business interest in an active, liquid retail market has no reason to frame the data any way other than constructively.