There is no sense of panic on the trading floor at 200 West Street. Not quite yet. Analysts move quickly between desks, coffee cups build up near terminals, and screens flicker in that familiar pattern of green and red. However, Goldman Sachs’ most recent market statement has a tone that goes beyond standard caution. It has memory.
The bank issued a warning in February 2026 that a fresh drop in U.S. stocks may lead to up to $80 billion in automated selling in a single month. Geopolitical shock or declining earnings would not be the culprit. It would be Commodity Trading Advisers — systematic funds programmed to react to price trends, not fundamentals, dumping stocks once technical thresholds are breached. This type of mechanical unraveling has a hauntingly familiar feel to it.
| Company | Goldman Sachs Group, Inc. |
|---|---|
| Founded | 1869 |
| Headquarters | 200 West Street, New York, NY, USA |
| CEO | David Solomon |
| Industry | Investment Banking, Asset Management, Financial Services |
| 2026 U.S. GDP Forecast | 2.6% Growth |
| Key Warning (Feb 2026) | Potential 10–15% Market Drawdown; $80B CTA-Driven Selloff Risk |
| Website | https://www.goldmansachs.com |
It seems that the purpose of this warning is structure rather than panic. CTAs don’t inquire about a business’s profitability. Whether AI revenue is increasing or margins are growing is irrelevant to them. They pay attention to cues. They sell when those indications switch. It has always been a little unnerving to see markets becoming more and more dominated by algorithms, much like watching traffic managed by unseen hands. Upon reversing momentum, liquidity may rapidly diminish.
Additionally, the bank noted that options markets are moving toward what traders refer to as “short gamma.” To put it simply, dealers who are in short gamma positions have to purchase into rallies and sell into drops, which intensifies rather than stabilizes swings. Usually, that configuration shows up right before turbulence. Similar traits were seen in late 2021: stretched valuations, muted volatility, and upbeat sentiment. The floor then moved.
It’s difficult to ignore how familiar everything sounds. David Solomon has made clear comparisons between the dotcom bubble of the late 1990s and the current AI-driven market boom. The analogy seems audacious, but not wholly inappropriate. Approximately one-third of the weight of the S&P 500 is currently held by a small group of tech behemoths known as the “Magnificent Seven.” Concentrated gains have pushed valuations higher and reduced market leadership.
Investors appear to think that those multiples will be justified by artificial intelligence. And perhaps it will. Data center building is growing, chip demand is strengthening, and corporate profitability linked to AI infrastructure have increased significantly. However, in several instances, expectations are rising more quickly than revenue. That disparity begs the question. It’s possible that the market is perfectly priced.
Outside brokerages in Midtown Manhattan, ordinary investors browse trading apps while riding the subway, talking about Nvidia and Microsoft with the same casual familiarity that was formerly reserved for Apple in 2007. The excitement is sincere. It feels brittle, too. Being optimistic about the internet in 2000 wasn’t incorrect; it was just too early.
Another dimension is added by Goldman’s caution regarding the “Great Software Stock Rout of 2026.” Already under pressure from growing rates and disruption from AI, many defunct software brands may see additional drops. The discourse surrounding software has changed from emphasizing expansion at all costs to questioning its longevity. Businesses that were once praised for their steady income are now being questioned about their competitive moats.
Meanwhile, interest rates continue to be high. Perhaps the most perplexing aspect of this cycle is that. High rates have historically compressed valuations. However, stocks are still trading at levels that indicate confidence rather than prudence. Whether investors are underestimating danger or are merely placing bets that economic growth will outpace monetary tightening is yet unknown.
In 2026, Goldman’s analysts predict 2.6% U.S. growth, which is barely recessionary. The economy appears stable on paper. Employment is still strong. Spending by consumers has not decreased. The balance sheets of corporations don’t show signs of trouble. However, markets don’t always adjust when economies contract. When placement gets crowded, they adjust.
The selloff scenario of $80 billion is not a prediction. It is a weakness. A technical unwind was brought on by violated levels on trading models that were humming softly in server farms throughout New Jersey, rather than by falling fundamentals. One feels a little uneasy after witnessing that dynamic in previous episodes: the flash crash, volatility spikes, and abrupt liquidity gaps.
The psychological layer is another. The AI trade is no longer merely a commercial one; it has become cultural. In their earnings calls, executives almost automatically bring up machine learning. Automation is marketed as a panacea by startups. The story is gripping. In times of invention, it always is. However, stories can become more inflated than profits and deviate from quantifiable fact.
Unaware of CTA placement or gamma exposure, travelers posed under enormous American flags this week while standing on the sidewalk close to the New York Stock Exchange. From the outside, markets seem to be quiet. Inside, traders discuss flows and convexity while modifying hedges in response to rising volatility.
Goldman does not foresee a collapse. Its long-term outlook is actually still positive. That detail is important. However, the company is indicating that the market’s infrastructure, including options positioning, algorithmic triggers, and concentrated gains, is being overextended in ways that are similar to previous pre-correction settings. As you see this play out, you get a déjà vu sense. Do not panic. Not sure. Only acknowledgment.
Warnings are often ignored by markets until they are no longer ignored. And the change can happen quickly, almost clinically, when technical selling and brittle sentiment collide. It’s unclear if this cycle will conclude with a slight 10% decline or something more drastic. However, Goldman’s ingredients—automated flows, crowded trades, and speculative enthusiasm—have already been seen. History also has a tendency to rhyme, even if it doesn’t repeat exactly.
