The Invesco QQQ Trust, which tracks the Nasdaq-100 index, ended a six-session losing streak Thursday morning after closing at $661.73 the day before. That 2.04% drop pushed the index into correction territory.
By mid-morning, QQQ traded at $678.38, up 2.52% from Wednesday’s close. The session range spanned from $661.89 to $683.68, a 3.3% swing in under three hours. Later readings showed the fund near $681.65, a 3.0% gain.
The Nasdaq-100 entered correction Wednesday, closing 9.8% below its June 2 record high of $745.34. At $678.38, the gap narrows to 9.0%—just enough to exit correction status if the day holds. The 10% threshold is a round-number convention with no fundamental signal behind it.
The fund’s concentration in a few large stocks means one earnings report can move it 2.5% in a single session. Microsoft’s 14.2% surge after reporting 43% Azure growth drove most of Thursday’s rebound. The broader market confirmed the move: the Nasdaq Composite rose 2.30%, while the Nasdaq-100 outperformed at +2.79%. The Dow added 0.60%, and the Russell 2000 lagged at 0.88%. The VIX fell 11.7% to 18.24, removing Fed-event premium rather than signaling a risk re-rating.
The trigger was the Federal Reserve’s decision to hold rates at 3.50%–3.75% on a 9–3 vote. Markets had priced 35% odds of an increase, so the hold itself wasn’t the issue. The bond market’s response was immediate: the 30-year Treasury yield surged 12 basis points to 5.21%, its highest since 2007. The 10-year rose above 4.67%, while the two-year fell four basis points—a bear steepener that hits QQQ’s long-duration holdings hardest.
QQQ’s top components commit $130 billion to $175 billion annually to infrastructure with returns expected years out. A 12-basis-point move at the long end isn’t background noise—it’s the discount rate applied directly to the thesis. The Fed chair’s framing amplified the effect by noting that tightening in financial markets is already doing part of the work.
The defining feature of Thursday’s session was that QQQ’s two most consequential earnings reports moved in opposite directions by a combined 23 percentage points. Microsoft surged 14.2% on a 43% Azure growth print. The broader tape confirmed the character of the move.
For an index fund, that dispersion creates a problem the wrapper can’t solve. A QQQ holder owns both sides of an argument they may have a view on and can’t express. The fund’s 2.5% gain is the net of a 14% winner and a 9% loser at comparable weights, plus a semiconductor complex rallying more than 7%. The case for index exposure: a stock-picker who owned Meta and not Microsoft had a catastrophic morning. The case against: the fund’s concentration means it isn’t diversifying away the risk it appears to be diversifying.
Related: WBD trial schedule July 31 with $7B fee
When five holdings account for 29.9% of the portfolio, and all are exposed to the same capital-intensity question, owning the index is a concentrated position dressed as a diversified one. That concentration has been building for years. As of July 28, Apple, Nvidia, Microsoft, Micron, and Amazon collectively accounted for 29.9%. Apple’s weighting stood at 8.42%, just ahead of Nvidia at 8.03%. In the sibling fund tracking the same index, the top ten holdings represent 45.91% of assets across 108 positions.
The index is modified market capitalization-weighted, with caps to prevent concentration from running further. Those caps exist precisely to prevent the current levels, but they haven’t stopped the drift. The market-wide context sharpens the picture: the top ten S&P 500 components now account for roughly 40% of total market value, above the 2000 peak. QQQ’s concentration sits well beyond that because its universe excludes financials and is drawn from a single exchange listing.
The practical consequence appears in the tape. A fund with 8.42% in Apple carries meaningful single-name event risk into tonight’s earnings, where options price a 3.8% move against a 1.6% average across the last four quarters. A fund with 4.15% in Amazon carries the same exposure to a report where the capital expenditure line matters more than the AWS growth line, and where options imply a 6% to 7% swing. Between those two positions, roughly 12.5% of the fund’s value is subject to two binary events after the close.
Concentration cuts both ways. It produced the run to $745.34 on June 2. It also produced the 11.2% drawdown into Wednesday. Anyone holding this fund should understand they’re not holding a diversified technology allocation—they’re holding a concentrated one with a hundred-name label.
The most underappreciated holding is Micron Technology, sitting at 4.77% in the sibling fund and inside the top five. A memory manufacturer carrying nearly 5% of a mega-cap growth index is unusual. Memory pricing is the marginal variable in the AI hardware supply chain, and Micron is the purest listed expression of it. When memory economics deteriorate, the stock moves violently, and at a 4.77% weight, that translates directly into index-level moves. Micron fell 9.9% during Wednesday’s rout. It climbed 7.4% on high volume Thursday.
That single position has been contributing roughly half a percentage point of index movement on days when it moves 10%. Combined with Nvidia at 8.03% and the broader semiconductor complex, the fund’s sensitivity to high-bandwidth memory pricing is far higher than its sector label implies. The Philadelphia Semiconductor Index is down roughly 25% since late June, after surging 87% in the second quarter—its best quarter on record. Applied Materials dropped 8.4% Wednesday, AMD lost 5.5%, Nvidia shed 3.5%. Thursday, the complex rebounded more than 7%, led by Lam Research at 18.3% on record quarterly revenue and earnings.
The June rebalance increased this exposure. When CoreWeave and four other AI-infrastructure names joined the Nasdaq-100, more than $800 billion of index-tracking money had to buy them. That’s the passive mechanism working as designed, adding AI-infrastructure beta to the fund at precisely the moment the theme peaked.
For a holder, the takeaway is that QQQ’s semiconductor exposure has become the dominant swing factor in both directions. Memory pricing—an input cost most equity investors never tracked—is now a first-order determinant of the fund’s monthly return. That’s not what most people think they’re buying.
According to the sponsor, the technology sector represented an average 67.29% share of QQQ throughout June. That figure should govern how anyone positions this fund inside a portfolio. Two-thirds in a single sector makes QQQ a sector fund with a broad-market name. It’s not a substitute for equity beta and it’s not a diversifier. Investors holding both an S&P 500 fund and QQQ own the same seven or eight companies twice, at higher aggregate weights than either fund discloses in isolation—and the correlation between them during drawdowns is close to one.
Related: Swiss franc gains ground against dollar
The tracking data confirms the profile. The sibling fund carries a five-year beta of 1.24 against the broad market, a trailing price-to-earnings ratio between 30.41 and 33.95, and a dividend yield of 0.43%. Year-to-date total return runs 8.02%, against 2.85% for its large-growth category. Over longer trailing windows, the fund has returned 17.27% against a category 29.03% on one measure and 20.73% against 23.49% on another—periods where concentration worked against it.
The index construction adds a wrinkle. The fund invests at least 90% of assets in index constituents, the underlying index is non-diversified by regulatory definition, and rebalancing occurs quarterly with annual reconstitution. That means concentration drift between rebalances is a feature: when a large holding runs, its weight rises until the next quarterly cap adjustment. The exclusion of financials is the other structural quirk. With the 30-year at 5.21% and a bear steepener rebuilding term premium, banks and insurers are among the few equity groups that benefit from higher long rates. QQQ owns none of them. That’s why the fund fell 2.1% Wednesday while the Dow lost 2.2% but the S&P 500 held within 3% of its record.
The disciplined framing is that this is a high-conviction growth and technology allocation with a 1.24 beta and two-thirds sector concentration. Size it accordingly.
The most striking market statistic of the week is the divergence between the two headline benchmarks. As of Tuesday, the S&P 500 sat less than 3% below its all-time high while the Nasdaq-100 briefly entered correction territory. That seven-point gap is entirely explained by composition. The S&P 500 carries financials, energy, healthcare, industrials, and staples alongside its technology weight. Those groups have been the beneficiaries of exactly the conditions punishing QQQ: energy has rallied with Brent near $89, financials benefit from a steeper curve, and defensives absorb rotation out of long-duration growth.
The divergence isn’t a technology earnings problem. Corporate results from the largest holdings have been strong—Microsoft beat on every line, Amazon’s AWS has accelerated three consecutive quarters, Apple is guiding to 14% to 17% revenue growth. What’s broken is the willingness to pay a high multiple for cash flows arriving in 2028, and that willingness is a function of the discount rate rather than the cash flows.
Layered on top is a positioning unwind that has nothing to do with fundamentals. Prime brokers including Goldman Sachs and JPMorgan have been issuing margin calls to hedge funds with concentrated AI exposure, most triggered automatically by contractual risk provisions. Goldman disclosed that roughly 16% of its prime brokerage exposure sat directly in AI memory-chip-related names as of June 30, and reported the largest cumulative increase in hedge fund gross leverage since it began tracking the data in 2016.
An AI-dedicated fund that scaled a $225 million seed into more than $20 billion is unwinding positions after steep losses, with its prime brokers marketing a block of its long and short holdings ahead of the open. Its portfolio spanned power generation, storage, and data center operators—sub-groups already under pressure. For QQQ, the implication is that the drawdown has been amplified by forced deleveraging in exactly the names it holds most heavily. That cuts both ways: short-side covering inside a marketed block adds to days like Thursday. Reading a 2.5% bounce as a fundamental all-clear misses where a meaningful part of it came from.
The fund’s sensitivity to memory pricing highlights how specialized risks now drive broad index returns.
Leave a Reply