NextFin News - US equities are staring at a severe one-sided risk: Bank of America strategists calculate that the systematic trading machines underpinning this year's rally have only about $9 billion of buying power left if stocks keep rising, but could be forced to dump as much as $163 billion if they fall. The warning lands as investors pull cash out of US stock funds at the fastest pace in months, wholesale inflation re-accelerates, Treasury yields climb to their highest levels since 2023, and the corporate buyback machine that has propped up the market enters its September blackout window. The question is no longer whether positioning is crowded; it is what happens when the market runs out of automatic buyers.
The Asymmetry: $9 Billion to Buy, $163 Billion to Sell
The numbers from Bank of America's Global Investment Strategy team sketch an unusually lopsided setup. In an up-market scenario, systematic strategies — commodity trading advisers, volatility-control funds and risk-parity vehicles — can add roughly $9 billion of net equity buying. In a flat market they flip into net sellers of about $1 billion. In a downturn, combined CTA and volatility-control selling could reach approximately $163 billion. That is an 18-to-1 ratio of potential selling to potential buying, and it is the cleanest expression of the fragility building under the surface of the rally.
The bank's estimates rest on assumed assets under management of about $300 billion for CTAs and roughly $200 billion each for risk-parity and volatility-control strategies. The mechanics are straightforward and unforgiving: these strategies size positions off realized and implied volatility. When volatility is low and falling, they buy automatically. When volatility rises, they sell automatically — without a view on earnings, margins or the economy. The asymmetry exists because they are already fully extended on the long side.
This is not a lone view. Deutsche Bank data shows volatility-control strategies have pushed equity allocations to the 100th percentile of their historical range, meaning there is essentially no room left to add exposure. Goldman Sachs puts global CTA net long equity exposure at about $146.5 billion, near the top of its own historical distribution and up roughly $2 billion from the prior week. When three different models built on different inputs — Bank of America's CTA flow model, Deutsche Bank's percentile ranking and Goldman Sachs' net-long estimate — all locate systematic equity exposure at or near its ceiling, the range of plausible outcomes narrows considerably.
Nor is the warning confined to positioning models. Nomura strategist Charlie McElligott has described the current volatility setup as "epic" in scale, warning that major trading losses may be about to surface. The cross-firm convergence is uncommon, and it is what elevates this from a routine strategist note to a market-structure warning.
The Money Is Already Leaving
Positioning stress is arriving alongside actual money exiting the market. Investors withdrew a net $11.12 billion from US equity funds in the week through September 2, following $22.72 billion of outflows the previous week, according to LSEG Lipper data. Large-cap funds led the exodus with $7.52 billion of net sales, while mid-cap and small-cap funds shed $572 million and $1.83 billion respectively. Sector funds reported net selling of $3.48 billion, led by technology, financials and industrials.
The destination of the fleeing capital is telling. Money market funds attracted nearly $48.76 billion in the same week — their strongest weekly intake in a month. Bond fund inflows fell to a five-week low of $4.27 billion, but short-to-intermediate government and Treasury funds remained popular. Investors are not exiting risk altogether; they are moving to the shortest end of the curve and to cash.
Bank of America's own weekly flow report, dated September 3, showed the same rotation in motion. Weekly flows totaled $30.0 billion into cash and $18.3 billion into bonds, while equities attracted just $2.8 billion — the smallest inflow in nine weeks. Technology funds shed $1.5 billion, the largest outflow since June, financials lost $0.9 billion for a fifth straight week, and China equities saw $5.3 billion of outflows. The bank's private-client book, which manages $4.7 trillion, recorded its biggest weekly equity outflow since October, and its equity ETF share count ticked down 0.1 percent for the week even as it remains up 6.9 percent for the year.
The geographic and thematic rotation is even starker over a longer window. Active managers are now most overweight Europe and most underweight the United States. Bank of America's note states it plainly:
"In the last three months, funds have cut exposure the most to USA."
Over the past 12 months, the artificial-intelligence theme has seen the steepest outflows of the eight global themes the bank tracks, at minus $93.3 billion, followed by quantum computing at minus $41.6 billion. Gold and rare earths have attracted the most buying. At the stock level, Meta has been the single biggest seller over the past year with $37.3 billion of cumulative outflows, while Microsoft, Samsung Electronics, JPMorgan Chase and Nvidia were among the month's most sold names.
The Third Pillar Is About to Crack
There is a third source of support, and the calendar is about to remove it. More than $1.1 trillion in announced buyback authorizations had returned to open trading windows through August, providing a reliable bid beneath US equities. But the pre-earnings blackout period begins to accelerate around September 12, according to Citadel Securities, and companies typically suspend repurchases for roughly four weeks before reporting results.
Blackouts do not create active selling. They remove passive support — and a market with exhausted systematic buying capacity is precisely the kind of market that can least afford to lose support. Scott Rubner, head of equity and equity derivatives strategy at Citadel Securities, has traced the imbalance to the speed of the July recovery: the market rose fast enough to consume the systematic buying fuel, leaving little left for the next leg.
The timing is the crux. The blackout acceleration, the September 11 consumer-price index release and the Federal Reserve's September 15-16 policy meeting all land within the same week. The producer-price report for August, released September 10, showed final-demand costs rising 0.4 percent for the month and 5.4 percent on an annual basis — above the 5.3 percent forecast and up from 4.8 percent in July, with energy prices jumping 4.2 percent as hostilities between the United States and Iran lifted oil above $100 a barrel. That print pushed the 10-year Treasury yield to its highest level since 2023, near 4.8 percent, and set the stage for the CPI number that arrives the morning this warning was published.
Why a Small Drop Can Become a Large One
The danger is not merely that systematic funds are crowded; it is that their reaction functions are mechanically one-sided and mutually reinforcing. A fall in prices lifts volatility. Higher volatility forces volatility-control funds to cut equity exposure. That selling pushes prices lower, which lifts volatility further, which triggers more selling — and CTAs, which trend-follow across equities and volatility, sell into both legs of the move simultaneously. With volatility-control allocations at the 100th percentile, the amount of equity exposure subject to mechanical unwinding is near a record, while the pool of systematic buyers that could absorb a dip is close to empty.
In market-structure terms, the shock absorbers have been used up during the rally. A market stabilized by algorithmic buyers works well in one direction: it amplifies gains, suppresses volatility, and encourages more leverage. The same machinery amplifies losses in the other direction. The 18-to-1 ratio is a measure of how far the market has to fall before the machines stop being buyers and start being sellers — and how much they will have to sell once they start.
The buyback blackout compounds this because it removes the one source of demand that is not volatility-sensitive. Corporate repurchases have been the most reliable bid in US equities for years, and their seasonal withdrawal leaves the mechanical players as the marginal market. When the marginal buyer is an algorithm that sells into rising volatility, the market's natural stabilizer becomes a destabilizer. That is a regime change in how the market functions, not a cyclical fluctuation.
Cyclical Shock or Structural Shift?
Is this a mean-reverting positioning squeeze or a durable change in who owns US equities? The evidence points to both, operating on different time horizons — and separating them matters, because getting the call wrong flips the conclusion.
The cyclical leg is clear and tradable. Volatility-control and CTA positioning is mean-reverting by design: these strategies buy back when volatility falls, and history shows the unwind tends to be fast and reversible. The 18-to-1 asymmetry is a function of where volatility sits today, not of any permanent change in corporate profitability. If the S&P 500 holds its ground and implied volatility drifts back toward the low teens, the same machines that would sell a dip become buyers again. A cyclical shock is painful but self-correcting.
The structural leg is more worrying and does not self-correct. The buyers of last resort have changed. The retail and passive flows that powered the post-2020 bull market have been supplanted by systematic strategies and corporate buybacks — both mechanical, both volatility-sensitive, both prone to withdrawing at the same moment. When the marginal buyer is an algorithm, the market's stabilizing mechanism becomes pro-cyclical. That shift does not reverse when volatility falls; it persists, and it raises the risk premium investors should demand for holding US equities through the next volatility episode.
There is also a valuation anchor that cuts against the doom case. The S&P 500 set an all-time high of 7,816.70 in August 2026 and, after four straight days of losses through September 10, closed at 7,591.79 — still roughly 15 percent above its level a year ago. Earnings from AI-related companies have stayed strong, and the economy has continued to grow. Crowding has been extreme for months, and the market has risen anyway. The bank's proprietary Bull & Bear Indicator has been flashing a sell signal since May 26 and still sits at 9.6, down only marginally from 9.7, in extreme bull territory. Since the signal triggered, the S&P 500 has added about 2 percent and global stocks about 1.9 percent. Crowding is a condition, not a catalyst — and conditions can persist longer than patience.
The Counter-Thesis: This Is Just Healthy Consolidation
The strongest argument against the warning is that crowded positioning is not new, and neither is September seasonality. Markets digest late-summer gains every year; buyback blackouts are a predictable calendar event, not a surprise; and volatility-control funds have been at the top of their range before without triggering a disorderly unwind. If the August CPI print lands near the 3.4 percent consensus and the Federal Reserve avoids a hawkish surprise at its September 15-16 meeting, yields could retreat from their multi-year highs, the dollar could stabilize, and the positioning overhang could be absorbed by fresh inflows rather than forced selling.
There is force in this view, but it rests on a single fragile assumption: volatility must stay low. The entire setup is stable only so long as nothing forces the mechanical sellers to act. A hot inflation number, a further oil spike, or a hawkish Fed shift would be enough to test it. Stability that depends on nothing happening is not stability; it is the absence of a test.
What Would Prove the Warning Wrong
The thesis has a clean falsifying signal. If the S&P 500 holds above 7,600 through the end of September while the Cboe Volatility Index closes below 14 for five consecutive sessions and the 10-year Treasury yield retreats below 4.5 percent, the positioning-overhang thesis is wrong — the market has absorbed the supply without a mechanical unwind. Conversely, a daily close below 7,400 with the VIX above 20 would confirm the feedback loop has begun. The VIX has already been climbing from its 2026 low near 14.13, closing at 16.34 on September 1 and 15.20 on September 2, before settling at 14.32 and 14.53 on September 3 and 4 — a sign that the market is already pricing a wider range of outcomes.
Outlook: Three Scenarios for the Month Ahead
Base case — contained volatility. The CPI print lands near expectations, the Fed holds or delivers a widely telegraphed move, and the 10-year yield stabilizes below 5 percent. Systematic funds trim rather than liquidate, the blackout window passes without incident, and the S&P 500 trades in a range between 7,400 and 7,700. The 18-to-1 asymmetry never gets tested.
Downside case — the unwind triggers. A hot inflation print or further oil spike pushes the 10-year yield above 5 percent and the VIX above 20. Systematic selling triggers, the blackout window removes buyback support, and the S&P 500 tests the 7,200 to 7,400 zone. This is the scenario Bank of America is flagging, and it is the one where the 18-to-1 ratio moves from a model output to a market reality.
Upside case — the asymmetry flips. Inflation cools, yields fall back below 4.5 percent, and the VIX drops toward 13. Systematic strategies turn from sellers into buyers, buybacks resume in October, and new highs follow. This path requires the mechanical players to be rewarded for patience rather than punished.
Across all three scenarios, the exposure is not evenly distributed. Quality balance sheets with visible cash flows are least vulnerable to multiple compression; long-duration growth names and the most crowded AI winners are most exposed. Sectors that have seen sustained inflows — healthcare and defensive areas of the market — are better positioned than technology and financials, where the outflows have concentrated. Bank of America's "Crowded Positives" screen flags Broadcom, TSMC, SK Hynix, ASML, CATL and UniCredit as heavily owned names with strong momentum, while its "Under-owned Negatives" screen highlights Coinbase Global, Blackstone, JD.com, Datadog, Mercedes-Benz and Saudi Aramco as stocks where weak momentum meets low fund ownership.
The takeaway is not that stocks must fall. It is that the market has run out of buyers who step in automatically — and in a market built on automatic buyers, that is the one condition that turns a routine pullback into a disorderly one. September has historically been the weakest month for US stocks; this year the seasonal headwind is arriving with volatility-control allocations at a record high, buybacks on pause, and the inflation calendar stacked in a single week. The asymmetry is the story. Whether it gets tested depends on the next inflation print.
Explore more exclusive insights at nextfin.ai.
