Fear vs. AI FOMO! Afraid of missing out and afraid of a crash at the same time? "Retrospective put options" become the new favorite hedge in the AI bubble.

After experiencing a record-breaking seven-week rise in the S&P 500 index, the US stock market suffered a significant pullback last Friday. However, this decline was not simply a profit-taking event - it was the result of three-fold pressure: reigniting inflation, a market structure threatened by bubbles, and quietly expanding structural vulnerabilities. The intertwining of these factors is forming a "perfect storm" unlike any seen before. Some investors believe that the rise in the US stock market driven by tech stocks appears to be a bubble, so they are turning to alternative options that can better withstand a potential crash. The "Fear-versus-AI FOMO" argument proposed by strategists at the end of 2025 still exists, with intraday volatility largely dependent on the next steps taken by President Trump. Last year, tariffs were the biggest concern, but now inflation has become the greatest threat, with the drop in the S&P 500 index last Friday attributed to the surge in US Treasury bond yields. US stocks plummet, rekindling concerns about valuation bubbles Inflation impact: Surging US Treasury bond yields are reshaping the pricing logic of risk assets Last Friday, the S&P 500 index fell 1.24% to close at 7,408.50 points, despite still recording its seventh consecutive week of gains, the "cracks beneath the surface" of the market are becoming increasingly visible. The Dow Jones and Nasdaq indexes also fell by 1.07% and 1.54% respectively, with the Semiconductor ETF dropping by 3.80% in a single day. NVIDIA plummeted by 4.42%, while Intel and Micron Technology both dropped by over 6%. The trigger for this decline was not trade policy or geopolitical tensions, but simultaneous selling in the global bond markets - the yield on the US 10-year Treasury bond climbed to 4.59%, its highest level since February 2025; the 30-year bond yield even reached 5.10%. This change in interest rates was not an isolated event. Globally, Japan's 30-year bond yield surpassed 4% for the first time in history, and the UK's 30-year gilt yield rose to its highest level in 28 years. Emmanuel Cau, Director of European Stock Strategy at Barclays Bank, described this as "the reigniting of inflation exacerbating the pressure on already fragile bond markets globally." The "global simultaneous surge" in bond yields is challenging the logic that "the AI bull market requires low interest rates" - the 30-year US bond yield has already surpassed 5%, the 10-year bond yield is approaching 4.6%, and 4.5% is widely considered the "danger zone" for the stock market. The deterioration in inflation data is a direct driver of the rise in interest rates. The US CPI reached 3.8% year-on-year in April, its hottest reading since 2023; the PPI soared to 6.0%, its fastest pace since 2022. Bank of America's Chief Investment Strategist, Hartnett, has issued a clear warning: if the monthly CPI rate of 0.4% continues, CPI could exceed 5% before the mid-term elections in November, he defines CPI exceeding 4% as "the land of the dragon" - historically, once inflation crosses this threshold, the S&P 500 has averaged a 4% decline in the following three months and a 7% decline in the next six months. The stubbornness of inflation is fundamentally changing the market's pricing of the Federal Reserve's policy path. The CME's "Fed Watch" tool shows that the market expects a 95% probability that the Fed will keep rates unchanged until July, and maintaining rates unchanged for the entire year has become a market consensus. Since Powell took office, there have been high hopes in the market for a rate cut in the short term - his preferred "tail mean PCE" inflation index is low, and a rate cut aligns with the political demands of the Trump administration. However, with stalled negotiations between the US and Iran, continuous high international oil prices, internal divisions within the Federal Reserve, and the new chairman's difficulty in quickly consolidating opinions to push for a rate cut. The CPI and PPI data released this week have both exceeded expectations, showing that energy costs are transmitting inflation to the consumer. As the Iran war impacts the oil market and drives up inflation, traders are increasing their bets, believing that the Fed may even hike rates as early as the beginning of 2027. For fervent followers of the AI market, this is a dangerous signal. Several buy-side fund managers have explicitly stated that the 30-year US bond yield remaining above 5% is "dangerous territory for the stock market" - a level that has already been reached. Benot Peloille, Chief Investment Officer at Natixis Wealth Management, bluntly warned that "while the stock market is still looking at the world through rose-colored glasses, interest rates are rising steadily." AI Mania: The Ghosts of Concentration, Valuation, and Bubbles If macro risks come from external threats, then structural problems within the market pose internal vulnerabilities. The core engine of this round of US stock market growth is highly concentrated. More than half of the gains in the S&P 500 index this year have come from just four stocks. According to Morgan Stanley's data, the top ten AI-related companies account for approximately 40% of the total market value of the S&P 500. The Philadelphia Semiconductor Index surged by over 70.5% in 2026, with Intel rising by 214.6%, far exceeding NVIDIA's previous gains. The valuation alarm has also been sounded across the board. The Shiller P/E ratio of the S&P 500 has climbed to 39.58, up nearly 13% from 35.08 a year ago, approaching the peak of 44.19 before the burst of the dot-com bubble in 2000. The concentration of the market value of the top ten components is around 40%, almost 50% higher than the 27% during the dot-com bubble era in 2000, surpassing the level of concentration risk seen then. The latest report from Bank of America strategist Michael Hartnett directly points to the crux of the issue: the current price of the Philadelphia Semiconductor Index is 62% higher than its 200-day moving average, exceeding the deviation seen before the "Black Monday" of 1987 and the Great Depression of 1929, and nearing historical extreme records before the bursting of the Mississippi Bubble in 1720. Hartnett plainly states that the current market is showing "index-level price movements, increasing market concentration, declining volatility, and stocks surpassing bond yields" - all typical signs of a bubble. However, there are fundamental differences from the dot-com bubble of 2000: the leading AI companies in this round all have strong profitability, with the forward P/E ratio of the tech sector as a whole around 30 times earnings, much lower than the 50 times seen during the dot-com bubble era. A report from CICC also points out that in terms of demand, investment intensity, and valuation, AI is not yet in a "classic bubble stage," but there is an "objective run on investment relative to demand." This divergence in views itself illustrates the issue: no one can deny the reality of the AI industry revolution, but what is truly worrisome is the way assets are being priced. Bears, led by "The Big Short" investor Michael Burry, directly compare the current surge to "the last few months before the crash of 2000" and reveal that they are shorting the semiconductor sector through put options. But contrarian investors present an intriguing counter-argument: unlike the widespread bullish sentiment in 1999, there are still many investors who remain skeptical - Bank of America data shows that global fund managers have reduced their equity allocation by two-thirds since March. Lurking Structural Bombs: Leveraged ETPs and the "Death Spiral," the market's "Synthetic Negative Gamma" Amplifier While tech stocks are being fervently pursued, a more covert risk is rapidly accumulating - the scale of leveraged exchange-traded products (ETPs) has expanded to historic levels. The latest calculations from Nomura Securities strategist Charlie McElligott are alarming: a 5% decline in the S&P 500 index in a single day could trigger passive selling totaling up to $187 billion from options traders, leveraged ETPs, and volatility control funds, creating a "death spiral" of "selling more as it falls," a scenario he describes as "jumping off a higher cliff." The operation mechanism of leveraged ETPs in a rising market is like "adding fuel to the fire" - additional stocks need to be bought at the end of each trading day to maintain the target leverage ratio, providing continuous fuel for the uptrend. However, the same mechanism operates in strict reverse in a declining market: funds are forced to sell massively to reduce exposure, which could trigger a stampede in illiquid conditions. The total size of leveraged ETPs has reached $179 billion, with 85% heavily concentrated in tech, AI, semiconductors, and related themes, and over the past month has seen net inflows of over $100 billion. Barclays Bank strategists' calculations show that the theoretical buying and selling pressure of leveraged funds on the S&P 500 index for every 1% fluctuation has surged from approximately $6 billion at the end of March to around $10.8 billion, and the impact of rebalancing funds during the closing session is becoming increasingly significant - its price effect is similar to the "short gamma" of options traders. In fact, when the AI concept stocks experienced a significant pullback in February 2026, the approximately $18 billion selling pressure from the rebalancing of leveraged ETPs was one of the key drivers of the decline, and the characteristics of a "technical stampede" were already evident. Strategies to Address "The Retroactive Put Option" Goes Popular: A Unique Weapon to Hedge Bubbles In such a dilemma - fearing a deep pullback, but also fearing to exit early and miss out on the AI boom - a more complex over-the-counter derivative has become a new favorite among institutional clients. This tool, known as the "retroactive put option," is designed to precisely address the current market dilemma: the exercise price is set based on the highest market price during the option's effective period, allowing investors to "look back" at historical highs to determine protection levels. Even if the market continues to rise and then crash, the protection level will rise with it, solving the problem of deep out-of-the-money failures that typical put options face due to fixed exercise prices. Neeraj Chaudhary, Director of Exotic Options and Liquidity at Bank of America for Europe, the Middle East, and Africa, has revealed to the media: "We have seen demand for retroactive put options from clients, as they want to hedge against potential rebounds in the market before a sell-off. Retroactive put options are very well-suited for this situation because their exercise price is set based on the highest index level within the term of the trade." To offset the higher cost of these options, Bank of America also recommends employing the strategy of "widening put option spreads" - by selling standard put options with lower exercise prices to partially raise funds for the purchase of retroactive put options. It is worth noting that the demand for retroactive put options is not new. As early as 2025, in an environment where US stocks were continuously hitting new highs, related over-the-counter exotic options products were already popular in the market. Now, with the accelerated rise of tech stocks, this hedging demand has heated up again, reflecting a deep-seated investor anxiety: the rise itself is becoming the biggest source of risk. The Evolution of Quantitative Strategies: From Enhanced Returns to Macro Defense Facing an increasingly complex environment, the role of quantitative investment strategies is also undergoing a significant change. Adrien Geliot, CEO of Premialab, stated: "After the outbreak of the Iran conflict, QIS (Quantitative Investment Strategies) gradually shifted its focus from enhancing returns to portfolio defense and macro adaptation. During the Iran crisis, the most effective systematic framework was able to adapt faster to the repricing of volatility, inflation expectations, and cross-asset trends than some subjective decision-making processes." However, not all quantitative strategies can effectively play a hedge role in market crises. Adrien Geliot, Head of Citi's Global Multi-Asset Structuring, warns that strategies that dynamically trade long VIX futures positions may put investors at a disadvantage when they truly need a long volatility exposure because "every crisis is different, and the triggers are also different." In the early days of the Iran conflict, the market experienced a severe repricing across assets - oil prices soared, inflation expectations jumped, and stock-bond correlations briefly turned positive, significantly reducing the diversification effect of a traditional 60/40 portfolio. During this period, the value of a self-adaptive strategy framework that can quickly adapt to changes in the macro environment was particularly highlighted. However, history has repeatedly shown that quantitative hedge strategies often perform well in backtesting, but actual trading results fall short of expectations due to factors such as unstable parameters and over-reliance on specific signals, a fundamental limitation that investors must be aware of when deploying such strategies.