Goldman Sachs Shifts AI Trade to Options, Urges Hedged Rebound Participation

Goldman Sachs has revised its artificial intelligence trading strategy, recommending investors "rent" upside participation through options rather than buying stocks outright, paired with put protection against further declines. The bank's derivatives desk said its momentum-factor equity basket has logged 30 single-day moves exceeding 5% this year, a frequency surpassing the dot-com bubble, the 2008 financial crisis and the COVID-19 pandemic. Despite earnings support and positioning near five-year lows that could lift equities into year-end, Goldman favors limited-loss option structures. It prefers March 2027 90% S&P 500 lookback puts at a cost of 1.82%, whose strike adjusts upward with daily market moves, removing the need to time entry. Among megacaps, Goldman cited an Amazon (AMZN) December 280/310 call spread, noting the stock is about 10% off its yearly high but holding support. Implied volatility for the Magnificent Seven sits in the 15th percentile of the past four years relative to the broader tech sector, while hedge funds and long-only institutions sold nearly $20 billion of tech, media and telecom shares recently, approaching July levels. For neoclouds, Goldman prefers CoreWeave (CRWV) convertible bonds over stock; the shares fell 30% this month while the converts slipped only 1-2 points on a delta-neutral basis, offering roughly 300 basis points of extra yield. JPMorgan warned of bankruptcy risk among neoclouds and possible bargain acquisitions by large tech firms. Goldman's Brian Garrett said the market may be in the "eighth inning" of mass deleveraging rather than the "second inning" of an AI ecosystem repricing. Confirming a bottom requires lower bond volatility, fading U.S. midterm election uncertainty and AI capital spending flowing into corporate earnings. Goldman also favors a South Korea ETF (EWY) call spread, betting on catch-up buying if the AI trade revives.