Futures
Access hundreds of perpetual contracts
TradFi
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Launchpad
Be early to the next big token project
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Even more intense than the DeepSeek storm! An article explaining: Where did the Wall Street "abandon AI" frenzy originate?
Since mid-2022, momentum traders have dominated the market with Sharpe ratios that would make Warren Buffett sweat. However, on Wednesday, everything nearly collapsed across the board—resulting in record-breaking losses.
In simple terms, momentum trading is “buying the recent winners and selling the worst performers.” According to industry statistics, Goldman Sachs’ “High Beta” or unconstrained momentum portfolio (GSPRHIMO) just experienced its worst day since 2022, surpassing the AI sell-off after the Deepseek storm in January last year. High Beta stocks are generally stocks with more volatile swings than the overall market.
If Tuesday’s market decline was driven by software stocks capitulating due to AI disruption risks, then Wednesday’s weakness was more focused on fundamentals and position adjustments, making the large-cap index movements appear deceptively stable compared to underlying turbulence.
This confirms Goldman Sachs analyst Lee Coppersmith’s findings from last weekend’s article, that beneath the calm surface of the US stock market, there are raging currents. Data shows that the average 1-week realized volatility of S&P 500 component stocks, relative to the overall S&P 500 volatility, just hit 6.88 last Thursday—placing it in the 99th percentile of 2023.
In other words, the average volatility of S&P 500 components is roughly seven times that of the index itself, which explains why some individual stocks have exhibited extreme movements in recent days, even though the S&P 500 appears relatively smooth on the surface, masking underlying turbulence.
What’s happening in the market?
So, what exactly occurred? Goldman Sachs found that the Thursday sell-off was mainly driven by severe underperformance of long strategies (“past winners” portfolio GSXUHMOM), which tend to favor cyclical, high-volatility, high-beta themes.
Echoing Goldman’s view, Morgan Stanley strategist Bryson Williams noted in a Thursday close report that the core feature of Thursday’s trading was an extreme factor reversal, with momentum suffering its worst retreat in over three years. Similar to Goldman’s unconstrained momentum portfolio, Morgan Stanley’s momentum portfolio experienced multiple standard deviations of decline, primarily driven by long liquidation rather than asymmetric deleveraging, as investors actively reduced their holdings of this year’s winners.
Market weakness was heavily concentrated in crowded themes that led earlier: such as AI beneficiaries, AI power, national security sectors, Bitcoin miners, and other high-beta, high-momentum sectors—while previously lagging sectors rebounded strongly across nearly all areas.
This “anti-AI” dynamic echoes Tuesday’s divergence and completely reverses market leadership: early-cycle stocks, chemicals, regional banks, and some large-cap defensive stocks performed well, as funds withdrew from previously consensus trades.
Notably, there was no obvious “trigger” for Wednesday’s moves. Goldman Sachs pointed out that the market was simply correcting from the pursuit of the strongest earnings expectations, with short-term technical indicators becoming overly extreme in the process.
More critically, the extremely fragile market structure amplified the volatility. Leveraged ETF rebalancing became a major driver, Morgan Stanley estimates that on Wednesday, leveraged ETFs faced about $18 billion in selling pressure—placing that day among the top ten in history. The selling pressure was mainly concentrated in Nasdaq, tech stocks, and semiconductors, with Nvidia, Tesla, AMD, Micron, Palantir, and other stocks experiencing significant selling.
Market makers’ gamma dynamics also failed to effectively ease the pressure: although market makers maintained long gamma positions in options, this exposure had been significantly reduced. After offsetting the impact of hedge strategies involving short gamma leveraged ETFs, the market essentially entered a net short gamma state, which caused selling pressure to accumulate and intra-day volatility to spike.
Wednesday’s market was particularly unusual in its disconnect between index and individual stock performance. Nearly three-quarters of stocks outperformed the S&P 500 during the day—an extremely rare historical phenomenon—while the index itself continued to decline, highlighting that pain was highly concentrated in specific sectors.
Additionally, retail participation was noticeably absent compared to previous trading days, contrasting sharply with the aggressive buying seen during last week’s rebound and yesterday’s software-led decline. Meanwhile, institutional selling remained stable but did not reach extreme percentiles. This further supports the view that this was a repositioning rather than forced liquidation.
Williams summarized, “Wednesday’s action clearly continued the rotation and dispersion trend that emerged on Tuesday—no longer limited to the software sector, but reflecting a broader ‘risk appetite reversal’ moment. Crowded momentum strategies are taking a heavy hit, although the decline at the index level remains manageable for now.”
Goldman Sachs also pointed out that sectors that performed strongly this year are experiencing sharp reversals: memory chips, rare earths, and unprofitable tech stocks have all given back most of their gains. This explains why all high-momentum portfolios saw declines on Wednesday, while themes with negative momentum rebounded across the board.
Given that Goldman Sachs Prime Book data shows net exposure at high levels (89th percentile over the past year), a larger correction in winning sectors could cause losses far exceeding typical sector rotations.
What’s next?
Goldman Sachs observed that on days like Deepseek storm Monday, post-COVID vaccine releases, or other high sigma days, the fundamental narratives for winners and losers both shift. In contrast, this Wednesday seems more like a result of rising volatility, stretched technicals, and theme divergence, with some high-growth themes’ outlooks unchanged.
Therefore, traders at Goldman believe that, based on historical experience, the kind of momentum stock correction seen on Wednesday often presents a good medium-term buying opportunity. However, aside from some fleeting moments on Tuesday and Wednesday, there are no signs of panic selling or complete capitulation. Given recent strong momentum and high positioning, short-term hedging remains reasonable.
Morgan Stanley’s trading desk shares a similar view: their short momentum basket rose 0.8% on Wednesday. Positions in crowded sectors were sold off: AI stocks fell 8.5%, storage stocks down 6%, and AI power stocks dropped 8%.
In summary, Morgan Stanley’s Williams said that everyone is asking why to sell (Williams disagrees that it’s due to AMD’s earnings—its results were actually very good). He suspects this may just be a VaR (Value at Risk) shock forcing some long positions to de-leverage.
Like Goldman Sachs, this Morgan Stanley strategist noted that this is often a good buying opportunity, but he remains cautious and believes that momentum and crowded long positions still carry downside risk… even if Thursday might see a short-term rebound.
(Source: Caixin)