The Worst of the Selloff May Be Over: How to Position for the Next Rebound？💹

English version：

The first thing I did this week was not look for a new trade. I closed the straddle I had put on Nvidia last week. The table below summarizes the positions before I flattened the trade.

The contrast was striking. The NVDA 227.5 call cost 4.39 and had fallen to 0.24, a loss of 94.53%. The put at the same strike and expiration cost 3.97 and had risen to 13.63, a gain of 243.61%. One leg was almost worthless, while the other had nearly tripled. Together they left the account with a profit of $2,530.50. That is the point of a straddle: you are betting on a large move, not on direction. Once the volatility has been realized, you should not overstay the trade.

$纳斯达克(.IXIC)$ $NQ100指数主连 2612(NQmain)$ $纳指三倍做多ETF(TQQQ)$ $微型NQ100指数主连 2612(MNQmain)$ $标普500波动率指数(VIX)$

I want to make this clear up front. I have already taken profit on the NVDA straddle I shared last time. If you are still holding it, pay attention to your exit. I also took profit on the call I bought near the lows last week. The same principle applies to a one-sided option trade: once it is profitable, it is better to close it while gamma is doing the most work. The longer you wait, the more time decay and a shrinking bid-ask spread can erode the gain. In the same account, the GLD spread returned 62.33% and the USO 138P/140P spread returned 83.22%, while the USO 160C/162C spread was down 153.42%. One screen, four very different outcomes. That is the nature of trading volatility.

$S&P 500(.SPX)$ $SPDR S&P 500 ETF Trust(SPY)$ $E-mini S&P 500 - main 2612(ESmain)$ $Cboe Volatility Index(VIX)$

The Worst of the Selloff May Be Behind Us

After closing those positions, I spent the weekend reviewing the market. One question matters most now: has the US equity selloff reached a bottom?

$Gold - main 2612(GCmain)$ $E-Micro Gold - main 2612(MGCmain)$ $1-Ounce Gold - main 2612(1OZmain)$ $SPDR Gold ETF(GLD)$ $iShares Silver Trust(SLV)$

I am watching three lines: Nvidia, the S&P 500, and the Nasdaq. My conclusion is that the sharp short-term decline may ease this week. If sentiment improves, the market could rebound. That creates an opportunity to sell a small amount of deeply out-of-the-money puts this week.

$Tradr 1.5X Short NVDA Daily ETF(NVDS)$ $NVIDIA(NVDA)$ $Micron Technology(MU)$ $Tesla Motors(TSLA)$ $Apple(AAPL)$

Start with the S&P 500. The key level on the chart is the 120-day moving average, currently near 7,597.18. It is also the lower edge of the previous consolidation range. Apply the same logic to Nvidia.

On the daily chart, the shares are trading inside a rising channel that runs from 189.80 to 243.37. After touching 243.37, the stock pulled back and is now near 214.The green horizontal line below the channel is at 194.53, which marks the base of the structure. The NVDA trade is therefore straightforward: I am betting that the stock will not break below 194 this week.

$China A50 Index - main 2610(CNmain)$ $Hang Seng Index - main 2610(HSImain)$ $Hang Seng Tech Index - Mar 2027(HTI2703)$ $Hang Seng Biotech Index - main 2610(HBImain)$

There is also a political reason for this view. Before the midterm elections, President Trump will want to keep the S&P 500 near current levels and avoid a sudden crash. Unless a funding shock or political accident occurs, the market is likely to remain relatively calm before the election. After the election, however, divided government is a strong possibility: a Republican president and Democratic control of both chambers. At that point, stocks could correct sharply as positioning risk and fears about political gridlock are released together.

Flows support that interpretation. The SOX semiconductor index fell more than 10 percentage points last week. Under normal circumstances, panic should have spread. Instead, highly leveraged bottom-fishing capital ignored the fear and bought SOXL, the triple-leveraged semiconductor ETF, producing its largest weekly inflow in years.

Leveraged ETFs attracted a net $9.5 billion during the week, the highest weekly inflow since records began in 2006 and far outside the historical range of volatility. Institutional positioning tells a different story. Goldman Sachs prime brokerage data show that hedge funds' net exposure to AI stocks has fallen to around zero.

$Philadelphia Semiconductor Index(SOX)$ $iShares Semiconductor ETF(SOXX)$ $Direxion Daily Semiconductors Bull 3x Shares(SOXL)$

In practical terms, most of the short-term selling may already be done. Unless a black swan arrives, the selloff will have difficulty accelerating from here.

Was last week's CPI report a black swan? I do not think so. CPI remained sticky and the broader tone was negative, but the most sensitive parts of the rates market behaved constructively. Two-year and 10-year Treasury yields failed to rise, instead breaking below key levels and extending their declines.

$Euro-Schatz - main 2612(FGBSmain)$ $2-YR T-NOTE - main 2612(ZTmain)$ $10-YR T-NOTE - main 2612(ZNmain)$ $iShares 20+ Year Treasury Bond ETF(TLT)$

Christopher Waller also did not increase the market-implied probability of an October rate increase. Markets still expect no hike in October. My judgment is that Treasury yields are near a short-term, staged peak. The selloff may therefore be approaching its final phase, and the S&P 500's current consolidation zone should be difficult to break decisively in the near term.

But This Is No Longer the Same Market

Do not confuse being near the end of the selloff with a return to clear skies. The rebound has begun, but the risk period is not over. The US market is no longer the market it used to be. The persistent rally that ignored every risk factor is unlikely to return over the next three to four months.

Selling in global AI-related stocks was heavy last week. More importantly, it appears to have moved beyond the usual institutional pattern of taking profit, buying the dip, and adding again. This episode looks more like an adjustment to institutional AI exposure. Net short positioning in AI is unlikely to become as large as it was before.

The numbers are clear. On Wednesday, October 14, the constituents of Goldman Sachs' global AI basket recorded their second-largest one-day net outflow in almost a year. The z-score fell to -2.8, exceeded only by the July reading. Long selling exceeded short selling by 3.5 to 1.

The longer-term picture is equally important. Goldman Sachs data on major institutional prime-brokerage positions show that net exposure to global AI stocks has fallen from roughly 33% in June to 26.1%.

This is a clear topping pattern. The probability that institutions will rebuild their positions aggressively has fallen. They may still sell high and buy low around current levels, but they are likely to keep overall exposure within a limited range.

The fundamentals are more troubling. Goldman Sachs estimates that hyperscaler capital spending will contribute about 11 percentage points to S&P 500 earnings growth in 2026, about 7 points in 2027, and a negative contribution in 2028 as depreciation catches up. The sequence is 11, 7, and -1. The direction is unmistakable.

That leads me to conclude that AI stocks may already have passed their peak. The marginal contribution of capital spending will fade, and that fundamental change will eventually be reflected in share prices. The only uncertainty is timing.

The bond market is already pricing the shift. The spread between yields on AI-related debt and non-AI corporate debt has widened steadily. At the start of the year both were near 200 basis points. AI spreads have since risen above 360 basis points, while non-AI BB and B spreads remain near 240 basis points. That gap is the price that the credit market is assigning to AI risk.

If AI stocks are still your main investment exposure, adjust your positioning promptly.

I am not calling for an immediate crash. The market is unlikely to collapse at once because any deterioration in the earnings of S&P 500 companies will take time to appear. Corporate interest expense as a share of debt remains far below a Treasury yield above 4%. It is only around 4%, and its long-term trend is still declining in line with the 10-year Treasury yield. The market can therefore continue to tolerate higher yields for now.

The Months Ahead Base Building and More Volatility

Put the evidence together and the likely path is straightforward. The market has more room to fall than to rise, but a major decline is not the most likely outcome before the midterm elections. The S&P 500 may build a base near current levels, bounce modestly toward its previous high or set a marginal new high, and then continue to trade sideways. After the election, US equities could experience a visible pullback. That is my base case, although the market will decide the details.

My Positioning for This Week

1. Nvidia. I will continue selling deeply out-of-the-money puts below 194 to collect weekly premium. If the stock breaks below 194, I will stop out immediately. This trade exchanges time value for income while imposing a firm risk limit.

2. Equity indexes. I will not trade them this week. Trades placed during a choppy period offer the least attractive odds.

3. Crude oil. Brent spot crude has completed a potential head-and-shoulders bottom, with a neckline near 109.

$WTI Crude Oil - main 2612(CLmain)$ $United States Oil Fund LP(USO)$ $Natural Gas - main 2611(NGmain)$ $E-Mini Natural Gas - main 2611(QGmain)$

An upside breakout would open more room to rise, but it would also increase volatility. I will therefore maintain the calendar spread from last week: long December WTI and short March WTI, seeking to profit from a bullish widening in the spread. I will stop out the entire spread if the front-month WTI price falls below its 10-day moving average.

There is an important technical point. These trades hedge one another. Only a stronger-than-expected rise in crude oil is likely to accompany a larger-than-expected decline in Nvidia. One side bets on oil strength, while the other assumes that Nvidia will not fall sharply. In effect, I am buying insurance on a single macro scenario.

If US equities suffer an unexpectedly sharp decline and S&P 500 futures break below their 120-day moving average, I will stop selling Nvidia puts and put on a QQQ straddle instead. I would buy an at-the-money call and put with the same expiration, seven days or more away, to capture a sharp rise in implied volatility.

This article reflects only my personal trading record and views and does not constitute investment advice.