Recent gains in semiconductor stocks have begun to waver, with internal warnings emerging from Goldman Sachs. On Wednesday, Goldman Sachs trader Thilo Deller noted that cracks are starting to appear in one of the stock market's most exuberant sectors, as semiconductor stocks have given back some of their recent advances. Concurrently, Goldman Sachs Managing Director Shawn Tuteja issued a briefing, stating that the market may be severely underestimating the systemic risk accumulating from leveraged ETFs in the semiconductor sector. In his briefing, Tuteja expressed a partially negative view on risk assets and highlighted a dynamic he believes the market has not fully priced. As chip stocks have surged dramatically, the scale of leveraged ETFs has expanded rapidly, accumulating a "short gamma" exposure that has reached a non-negligible magnitude. A reversal in the trend could trigger a chain reaction of forced selling through mechanical deleveraging. The immediate catalyst for the semiconductor stock pullback came from a policy statement in South Korea. A senior South Korean policy official publicly stated that the country should design a national dividend system to return a portion of AI excess profits to the public to avoid a K-shaped economy exacerbating social inequality. This statement quickly sparked market concerns about the future direction of AI industry profit distribution policies, leading to a decline in chip stocks. However, the official later clarified that the suggestion was intended to utilize excess tax revenue, not corporate profits, to support redistribution policies. Despite this clarification, sentiment in the semiconductor sector was already unsettled, leading to a partial retracement of recent gains. Goldman Sachs data reveals a deeper structural concern. Since the end of March this year, the Philadelphia Semiconductor Index (SOXX) has gained over 70%, and the assets under management in leveraged products tracking individual semiconductor stocks and semiconductor ETFs have surged sharply. According to Goldman Sachs estimates, the combined long semiconductor exposure held by these leveraged products is close to $100 billion. The operational mechanism of such products inherently places them in a "short gamma" position—requiring them to buy more when the underlying asset rises and forcing them to sell when it falls to maintain their target leverage ratio. Tuteja estimates that for the semiconductor sector alone, the daily dollar gamma exposure these leveraged products need to reset is approximately $2 billion. In other words, if the semiconductor sector rises 1% in a day, the leveraged ETF system requires a net purchase of about $2 billion; if it falls 1%, it requires a net sale of about $2 billion. Notably, the scale of this short gamma has nearly doubled over the past 6 to 9 months. Tuteja pointed out that compared to S&P 500 index options, semiconductor option pricing exhibits a significant anomaly: over the past year, the cost of holding semiconductor options has been notably lower than that of S&P 500 options. This divergence has become increasingly pronounced, especially in recent months as leveraged ETF scale has expanded rapidly and spot prices have continued to break higher. More alarmingly, the pricing of the left tail in semiconductor options appears to completely ignore the risk of a single-day gap down in a deleveraging scenario. Goldman Sachs regression analysis shows that, using SMH (the semiconductor ETF) as a reference, the price of 1-month 10-delta out-of-the-money put options relative to at-the-money options is at historically low levels. Similarly, regressing the price of 1-month at-the-money options against three times the price of 10-delta out-of-the-money put options also indicates that tail option pricing is low—an underestimation that is particularly prominent given the current scale of the leverage system. Tuteja's conclusion is that once semiconductor spot prices decline, leveraged products will be forced into mechanical deleveraging. The speed of deleveraging will depend on daily volatility; the more severe the volatility, the larger the single-day deleveraging scale. The current options market is clearly underpricing this gap risk. Based on this analysis, Tuteja recommends investors consider purchasing tail hedges in the semiconductor and AI sectors to manage potential risks arising from the accumulation of leverage within the system, especially in an environment where downside implied correlation does not fully price in gap risk. Goldman Sachs specifically favors holding downside options on SMH and downside protection on its AI leader stock basket, GSTMTAIP. Additionally, Goldman Sachs has observed some clients opting to build custom baskets of semiconductor stocks. Since the implied volatility of such custom baskets is typically lower than that of SMH, the holding cost is lower during calm market periods. However, in the event of a deleveraging event, their correlation with SMH is expected to remain highly synchronized, thereby achieving a hedging effect.
Comments