The February 2026 Selloff: Anatomy of a Multi-Trillion Dollar Wipeout

Between February 3 and February 5, 2026, a cascade of selling erased roughly $1 trillion from software stocks alone, crashed Bitcoin nearly 50% from its October peak, and sent silver plunging 31% in a single session, its worst day since the Hunt Brothers' corner collapsed in 1980. Gold, which had touched $5,600 an ounce days earlier, shed 11% in a single Friday afternoon. The S&P 500, the Nasdaq, crypto, precious metals, and emerging markets all fell together, while the VIX jumped above 22 and $83 billion poured into money market funds in a single week. There was no Lehman moment, no pandemic. A series of anxieties, filtered through leverage, algorithms, and gamma hedging, produced selling far more violent than any single cause warranted.

Bull markets form over years of compounding earnings, expanding multiples, and gradually increasing risk appetite. Bear markets arrive in days. The math guarantees it, the microstructure amplifies it, and the psychology of fear makes it happen faster than anyone expects. Understanding why requires pulling apart the machinery: the algebra of percentage losses, the carry trades, and the zero-day options that now function as the market's nervous system.

The February 2026 selloff did not begin with stocks. Gold moved first.

On January 30, President Trump nominated Kevin Warsh as the next Federal Reserve Chair. Warsh, a former Fed governor known for hawkish views on monetary policy, was the worst possible pick for anyone positioned in "anti-fiat" trades. Gold had been on a long run to nearly $5,600 per ounce. Silver had blown past $120. On January 31, gold fell roughly 11% to $4,745, its steepest intraday drop in over four decades. Silver fell 31% from $122 to $78.53. The CME Group hiked margin requirements 36% for silver futures and 33% for gold, forcing leveraged traders to liquidate positions they could no longer hold. An estimated $15 trillion in paper wealth disappeared from precious metals markets in two trading sessions.

Then came AI. On January 31, Anthropic released Claude Cowork, a suite of AI agent plugins designed to handle legal research, sales workflows, marketing analysis, and data processing. These were not chatbots offering suggestions. They were tools positioned as replacements for the enterprise software that Salesforce, ServiceNow, and Intuit had spent decades building. On Tuesday, February 3, when markets reopened, the software sector took the hit. Goldman Sachs' basket of U.S. software stocks fell 6% in a single session, their worst day since the April 2025 tariff crash. ServiceNow dropped 7.6%. Salesforce fell 7%. Intuit lost 11%. Thomson Reuters plunged 16%, its worst day on record. LegalZoom fell nearly 20%. Software, financial services, and asset management stocks lost an estimated $285 billion in market value on Tuesday alone. For the first time, markets treated AI not as a tide lifting all boats but as a zero-sum force displacing incumbents.

Selling continued through Wednesday and Thursday. On February 4, the ADP employment report showed only 22,000 private jobs created in January, a figure that would have been negative without a surge in education and health care hiring. The next day, the JOLTS survey showed job openings at 6.54 million, the lowest since September 2020, with the ratio of openings to unemployed workers falling below 1.0 for the first time since the pandemic. AMD reported a weak forecast and fell 17%. Alphabet announced $175 to $185 billion in 2026 capital expenditure, well above estimates. Amazon followed after the bell with a $200 billion capex figure that sent its stock down more than 10% in after-hours trading.

By the close on February 5, the S&P 500 sat at 6,798, down roughly 2.5% from its recent high and negative for the year. The Nasdaq Composite had shed about 4.5% over three days, its worst stretch since April. The Russell 2000 dropped roughly 2% in a single session. South Korea's KOSPI fell 3.86%. India's Nifty IT Index had its steepest intraday decline since March 2020. Bitcoin, which had peaked above $126,000 in October, crashed through $70,000, then $65,000, briefly touching $60,033, its lowest price in over a year. Exchanges liquidated more than $1 billion in leveraged crypto positions in 24 hours. The iShares software ETF had lost nearly $1 trillion in market value in seven trading days and sat 28% below its recent high.

The only things that went up were Treasuries, money market funds, and the VIX.

The most basic reason selloffs destroy wealth faster than rallies create it is arithmetic. A stock that falls 50% must rise 100% to get back to where it started. A 33% decline requires a 50% recovery. A 75% loss needs a 300% gain. This is just how multiplicative returns work.

The compounding trap makes this worse over time. A $100 stock that suffers a negative 50% compound annual growth rate for five years falls to $3.13. If it then achieves a positive 50% CAGR for the next five years, it only recovers to $23.73. Not $100. Not even close. The loss and the gain are nominally identical, but the loss came first and hollowed out the base from which gains compound.

The historical record backs this up. The average bear market since 1928 has lasted 289 days and declined 35%. The average bull market has lasted 988 days and gained 112%. Bears move roughly three times faster per unit of decline than bulls move per unit of gain. The COVID crash took 33 calendar days to erase 34% of the S&P 500's value. The Nasdaq's 83% drawdown after the dot-com bubble took 15 years and 5,442 trading days to fully recover. Even among the 20 best-performing individual stocks over the past four decades, the median maximum drawdown was 72%, with a median peak-to-trough duration of nearly three years and a median recovery time of over four years. Winning stocks suffer catastrophic drawdowns too. They come back. Most stocks don't.

The math has a psychological counterpart. Prospect theory established that humans feel losses roughly twice as intensely as equivalent gains. When a portfolio drops 30% in a week, nobody calmly calculates that they need a 42.9% gain to recover. The response is panic and the urge to sell. That behavioral response turns the arithmetic asymmetry into a self-reinforcing cycle: losses trigger selling, selling creates more losses, and the cycle runs until either prices fall far enough for new buyers to step in or a central bank intervenes.

The math explains why losses are harder to recover from. The market's infrastructure explains why they happen so fast. The February 2026 selloff was not driven by millions of individual investors independently deciding to sell. Interlocking mechanical systems drove it, each responding to price signals in ways that amplified the decline.

Start with volatility targeting and risk parity. An estimated $2 trillion globally sits in volatility-targeting strategies, with another $300 billion in risk parity funds. These strategies target a set portfolio volatility. When realized volatility is low, they add leverage and equity exposure. When volatility spikes, they sell to reduce risk. During March 2020, a stylized risk parity portfolio had to sell assets worth nearly 225% of its capital to meet its volatility target. During the August 2024 selloff, estimates suggested systematic strategies would dump $70 to $80 billion in equities in a single day, with at least $90 billion more to follow over subsequent sessions. These are automatic responses hardwired into code.

Then there is the gamma trap. The options market has become a major amplifier of equity moves. Over 50% of S&P 500 options volume now comes from same-day expiry contracts (0DTE options that barely existed before 2022). These contracts carry extreme gamma: small movements in the underlying index create enormous shifts in the delta that dealers must hedge. When the market sits in "positive gamma" territory, dealers are naturally long options, and their hedging dampens volatility, buying dips and selling rallies. When the market enters "negative gamma" territory, the dynamic reverses. Dealers who are short options must sell as markets fall and buy as they rise, amplifying moves in both directions. SPX options carry approximately $80 billion in gross gamma exposure. A 1% move in the index translates to a delta shift of roughly $80 billion, all of which must be hedged. When that hedging is procyclical, prices fall, dealers sell to hedge, prices fall further, dealers sell more.

The interaction between gamma, vanna, and charm makes this worse during selloffs. Vanna measures how an option's delta changes with implied volatility. When volatility surges, the deltas on short put positions climb, forcing dealers to sell more of the underlying asset. Charm, which measures how delta changes with time, normally provides a gentle buying pressure as out-of-the-money puts decay. During sharp selloffs, charm flows are overwhelmed by gamma and vanna effects, and the options market stops absorbing shocks and starts causing them.

Margin calls add fuel. As of December 2025, FINRA-reported margin debt stood at $1.23 trillion, a seventh consecutive record high. Relative to GDP, margin debt hit 3.91%, well above both the dot-com bubble peak of 2.6% and the 2007 pre-crisis peak of 2.5%. When prices fall, portfolios dip below maintenance margin requirements, brokers issue margin calls, and investors must deposit cash or liquidate. That forced selling drives prices lower, which triggers more margin calls, which forces more selling. Research using Indian stock market data found margin trading creates "excess comovement" during crises, meaning stocks fall together more than fundamentals warrant, driven largely by brokers liquidating concentrated client positions. This comovement effect appears only during selloffs, never during rallies. There is no equivalent forced-buying mechanism that creates upward cascades.

Finally, algorithmic trend-following. CTA and managed futures strategies, managing roughly $318 billion globally, follow price trends across asset classes. When prices breach technical levels, these algorithms sell. During February 2026, different systematic strategies deleveraged at different speeds: momentum strategies reacted within hours, while volatility-targeting funds adjusted over days as trailing realized volatility calculations incorporated the new data. The staggered nature of these unwinds means systematic selling persists for days or weeks after the initial shock, producing the "slow bleed" pattern that characterized February.

The combined effect is sometimes called a "doom loop": dealer hedging flows trigger CTA momentum signals, which propagate negative price action, which causes volatility-targeting funds to deleverage, which increases realized volatility, which forces risk parity unwinds, which pushes cross-asset correlations toward one. Each mechanism feeds into the next. Each was designed independently and rationally. Together, they produce synchronized selling across asset classes.

When headlines announce that trillions in market value have been "wiped out," the obvious question is: where did the money go? The answer matters for understanding what a selloff actually is.

Most of the wealth that disappears during a crash never existed as cash. Market capitalization is a marginal pricing concept, not a measure of money flowing through the system. If Apple has 15 billion shares outstanding and the last trade occurs at $200, Apple's market cap is $3 trillion. But only a tiny fraction of those shares traded at $200. The vast majority sit in retirement accounts, index funds, and insider holdings, untouched. If the next trade happens at $190, Apple's market cap drops $150 billion. But only one trade, maybe for a few thousand shares, took place. The $150 billion was a mark-to-market revaluation, not a transfer of funds. The analogy to housing is useful: if three houses on a street of twenty sell for $400,000, the street's "market cap" rises by $2 million, but only $1.2 million in cash changed hands. In a selloff, the reverse happens. A small volume of desperate sellers reprices every outstanding share.

During the February 2026 selloff, total U.S. equity trading volume on a given day might have been $500 to $800 billion. But the market cap decline across U.S. equities was several trillion. The difference is the leverage of marginal pricing. For every seller, there was a buyer. The cash that the seller received came from the buyer. What changed was the price at which transactions cleared, and because that price sets the mark for every outstanding share, the aggregate market cap shifted by multiples of the actual money that changed hands. Net selling, meaning the excess of selling pressure over buying that pushed prices down, was perhaps tens of billions. The trillions in "destroyed" value were unrealized gains that simply ceased to exist at the new marginal price.

Real money did move, though, and it moved toward safety. In the week through February 4, $83.09 billion flowed into U.S. money market funds in net purchases, the largest weekly inflow since December 2025. Short-to-intermediate investment-grade bond funds attracted $6.34 billion, their largest weekly inflow since at least 2022. Technology sector funds lost $2.34 billion in outflows. Total money market fund assets stood at $7.71 trillion as of late January 2026, with projections of another $500 billion in inflows through the year. Money market yields of 3.5% to 3.7% made cash an increasingly rational alternative to equities losing 5% in an afternoon.

This pattern repeats. During March 2020, government money market funds absorbed $834 billion in a single month. During 2022's bear market, money market funds took in $142 billion in the fourth quarter alone. The hierarchy of capital flow during selloffs is stable: first into money market funds, then into Treasury bills and short-duration bonds, then into bank deposits, and finally into sector rotation within equities. In February 2026, even as technology lost $2.34 billion, industrial funds gained $2.11 billion and metals and mining attracted $1.44 billion. Not all equity money left. Much of it rotated.

The vast majority of the headline losses were never "money" at all. They were the gap between the price someone last paid for a share and the price the market now assigns to it. When that gap closes, the wealth doesn't flow anywhere. It stops existing.

The February 2026 selloff was global. South Korea's KOSPI fell 5.26% on February 2 and then 3.86% on February 5. India's Nifty IT Index had its worst day since March 2020. Japanese, Australian, European, and emerging market equities all fell. Bitcoin, supposedly uncorrelated with anything, crashed alongside software stocks and precious metals. The old saying that "in a crisis, all correlations go to one" played out in real time.

This was not always the case. In the 1990s, the Asian financial crisis remained largely regional for months. The 1987 crash took hours to propagate across time zones. Today, contagion is nearly instantaneous, and the reasons are structural.

The algorithmic web is a big part of it. Fundamental discretionary traders account for only about 10% of equity trading volume, with roughly 60% driven by passive and quantitative strategies. Some estimates put the machine-driven share of daily U.S. equity moves at 80%. These algorithms operate across markets and time zones simultaneously. When S&P 500 futures sell off during Asian hours, algorithms in Tokyo, Hong Kong, and Sydney respond within milliseconds, selling local index futures and triggering cascading sales in underlying equities. CTA trend-following strategies react to price moves regardless of which exchange or time zone generated them. No human is in the loop to say "this is a U.S. software problem, not a Korean semiconductor problem." The algorithm sees a price signal and acts.

ETF arbitrage is another transmission mechanism. U.S. ETFs account for over 31% of total equity trading and hold more than $6.2 trillion in assets. When a country ETF trades at a discount to its net asset value, authorized participants buy the cheap ETF shares, redeem them for the underlying securities, and sell those securities. This arbitrage, designed to keep ETF prices aligned with their components, becomes a direct channel for transmitting U.S. selling pressure into foreign markets. Research across 41 countries found that an increase in the U.S. VIX directly creates selling pressure in country ETFs, which the arbitrage mechanism then transmits to local markets. The introduction of a country ETF itself increases that country's stock market correlation with the U.S., especially during high-volatility periods. ETFs have linked the world's equity markets into a single system, with the VIX as the driver.

The largest source of global contagion, though, is leverage that spans borders and asset classes. The yen carry trade, in which investors borrow cheaply in yen and invest in higher-yielding assets elsewhere, was estimated at roughly ¥40 trillion ($250 billion) before its unwind in August 2024. When the Bank of Japan unexpectedly raised rates and the yen surged 6% in a week, Japanese investors and hedge funds who were short yen sold their most appreciated assets: U.S. momentum stocks. This is why Nasdaq performance became correlated with a Bank of Japan interest rate decision. The Nikkei fell 12.4% on August 5, 2024, its worst day since 1987. High-yielding currencies from Mexico to Brazil to South Africa fell sharply. The carry trade unwind showed that leverage denominated in one currency transmits shocks to equity markets on the other side of the planet.

The Treasury basis trade carries a similar risk. Hedge funds exploiting the tiny price gap between cash Treasuries and futures contracts currently hold net short futures positions exceeding $800 billion, using leverage of 20x to 50x. In March 2020, when margins spiked and repo funding dried up, basis traders sold roughly $100 billion in Treasuries in days, driving yields higher at the exact moment investors expected bonds to rally. The Fed had to purchase $1.6 trillion in bonds to prevent a systemic collapse. The trade has only grown since.

The dollar itself creates another feedback loop. The U.S. dollar consistently strengthens during global selloffs. During March 2020, the dollar index surged from 96 to 103, causing severe dollar shortages worldwide. Countries with dollar-denominated debt face rising servicing costs. Their central banks sell dollar reserves to defend local currencies, but those reserve sales often mean selling U.S. Treasuries, which amplifies stress in the very market investors are fleeing to for safety. The U.S. net international investment position stood at negative $27.6 trillion as of the third quarter of 2025, meaning foreigners hold far more American assets than Americans hold abroad. When those foreign holders need to raise dollars, they sell their most liquid U.S. positions, adding to the selling pressure.

In February 2026, the contagion path followed a familiar pattern. Bitcoin's crash triggered margin calls that forced traders to dump precious metals. Gold's 12% single-day crash generated further margin calls across multi-asset portfolios. Software selling in New York transmitted to RELX, Thomson Reuters, and SAP in London and Frankfurt. South Korean and Taiwanese semiconductor stocks fell in sympathy with AMD and Qualcomm. The fact that these assets have no fundamental relationship to each other was irrelevant. They were connected by leverage, by algorithmic cross-referencing, and by the margin accounts of investors who held all of them.

Not everyone loses money in a selloff. Some market participants depend on volatility for their returns, and early 2026 was good to them.

Market makers did well. Citadel Securities posted $3.4 billion in net trading revenue in the first quarter of 2025, up 45% year over year, with net income at $1.7 billion and an EBITDA margin of 58%. The firm distributed approximately $5 billion in profits to investors across 2025. Jane Street reported $20.5 billion in trading revenue for 2024, with projections north of $7 billion for a single quarter in 2025. Virtu Financial's market-making division saw net trading income rise 41% to $1.2 billion in 2025, with fourth-quarter earnings beating analyst expectations by 43%. These firms profit from wider bid-ask spreads and higher trading volumes, both of which surge during selloffs. Citadel's algorithms execute millions of trades per second, capturing the spread between sellers and buyers. During the April 2025 tariff volatility, Citadel's daily Treasury volumes hit $70 million in per-basis-point sensitivity, three times their COVID-era levels. Volatility is their product.

Short sellers collected $24 billion in paper gains from software stock shorts in the first five weeks of 2026. Software and AI-related stocks had fallen roughly 20% since January 1, and hedge funds were increasing their short positions rather than covering. Microsoft's short interest climbed approximately 20% in a single week. The most heavily shorted names included TeraWulf (over 35% of float), Asana (25%), and Dropbox (19%). Microsoft, historically a "reversal stock" where shorts cover on declines, was now trading like a momentum-driven, distressed name, with shorts increasing into weakness. Institutional investors saw the software selloff as structural, not a buying opportunity.

CTA and managed futures strategies performed as designed. The Simplify Managed Futures Strategy ETF rose 1.43% on February 4 while the Nasdaq fell 1.5%. One prominent trend-following fund surged 45.62% in February 2024 during a different market dislocation. CTAs do well with sustained directional moves and cross-asset volatility, which is exactly what early February 2026 delivered.

Put option holders collected too. When stocks fall and volatility rises at the same time, put options benefit from both delta gains (the stock moving in your favor) and vega gains (implied volatility expansion increasing the option's value). A software-sector put bought when the VIX was at 17 doubled or tripled in value as the VIX jumped above 22 and underlying stocks fell 7% to 20% in single sessions. PayPal's 20%-plus single-day collapse on February 4 turned modestly priced puts into big winners.

On the other side of these trades sat the losers, and the common thread was leverage. At $1.23 trillion in outstanding margin debt (3.91% of GDP, exceeding both the dot-com and pre-financial-crisis peaks by more than 50% in relative terms), the U.S. equity market was running on borrowed money to a degree never seen before. Any meaningful decline triggered the cascading margin call dynamic described above. Retail investors, who now account for 23% to 36% of U.S. equity market order flow depending on the day, had poured $2 billion into leveraged Nvidia ETFs in a single session during the DeepSeek selloff in late January. Many of those positions were deeply underwater by February 5. Estimates suggested retail portfolios had declined roughly 10% from their peak, underperforming the S&P 500's 8.5% drawdown, meaning retail investors were concentrated in the riskiest names.

The dip-buying behavior present in prior selloffs was absent from software stocks. Retail investors bought $430 million in broad S&P ETFs during one session, but the sector-specific buying that had previously arrested software declines did not materialize. Something had shifted in the market's psychology. For the first time, the narrative around AI shifted from "this will make everything more valuable" to "this will destroy incumbent business models," and buyers were unwilling to step in front of that narrative shift.

The wealthiest people are often among the biggest losers. Larry Ellison, Oracle's co-founder, entered 2026 worth approximately $247 billion. By February 5, he had lost $49 billion, roughly 20% of his fortune, making him the biggest loser on the Bloomberg Billionaires Index for the year. He had already lost $22.6 billion in a single day during the DeepSeek selloff on January 27 and $24.9 billion in a single day in December 2025 when Oracle fell on AI capex concerns. Ellison owns approximately 41% of Oracle. Selling would mean losing control of his company, triggering billions in capital gains taxes, signaling weakness to the market, and potentially accelerating the decline he'd be trying to escape.

The software billionaires as a group lost more than $62 billion in the first five weeks of 2026. Adam Foroughi, the CEO of AppLovin, saw his net worth drop from $27.3 billion to $17.3 billion, a 30% decline, as his company's stock fell 42% year to date under the combined weight of a short-seller report, AI disruption fears, and the broader software rout. His two co-founders suffered similar percentage losses. Dave Duffield, co-founder of Workday, fell 19% to $11.3 billion as the stock hit a three-year low. Scott Cook, founder of Intuit, dropped 17% and fell off the Bloomberg 500 richest list entirely.

Jeff Bezos lost $9.21 billion on February 5 alone, before Amazon's after-hours plunge on its $200 billion capex announcement potentially added billions more to the damage. Jensen Huang of Nvidia lost roughly $5 billion in a single session, having already suffered a $20.1 billion single-day loss during the DeepSeek crash, the biggest single-day loss of value for any public company in history.

The "billionaire problem" is a concentration problem. These founders and executives hold enormous stakes in single companies for reasons that make diversification practically impossible. Selling means losing control: Ellison's 41% stake gives him effective control of Oracle, and meaningful selling would dilute that and invite activist investors or hostile boards. Selling triggers taxes: under the current U.S. tax code, unrealized gains are not taxed, but any sale crystallizes a capital gains liability approaching 25% or more of the proceeds. The rational strategy, which most billionaires follow, is to borrow against their stock at low interest rates, spend the loan proceeds, and never sell. This is the "buy-borrow-die" approach: borrow against appreciated stock, live on the proceeds, and when you die, your heirs receive a stepped-up cost basis and never pay taxes on the original gains.

Selling also sends a signal. When a founder-CEO sells stock, the market reads it as a lack of confidence. The resulting price decline is often self-reinforcing, punishing the seller for the act of selling. And the stock that cannot be sold often serves as collateral for billions in personal borrowing. Elon Musk, who describes himself as "cash poor" despite a net worth exceeding $660 billion, has used Tesla shares as collateral for personal loans. A significant decline in the stock would trigger personal margin calls, forcing sales at the worst possible time. The concentrated billionaire is stuck: too wealthy to need to sell, too concentrated to survive a crash unscathed, too constrained by taxes and control and collateral to diversify.

Mark Zuckerberg lost $29.8 billion in a single day in February 2022 when Meta reported its first-ever decline in daily users and the stock fell 26%. Musk lost $35 billion in November 2021 after tweeting about selling 10% of his Tesla stake. These are the wealthiest people alive, losing more in a day than many countries' annual budgets, because their wealth is concentrated in a single volatile security they cannot sell.

Modern markets create an illusion of liquidity that holds up when things are calm and breaks down when they aren't.

Before decimalization and the rise of high-frequency trading, the New York Stock Exchange relied on human specialists with formal obligations to maintain orderly markets, providing liquidity even during stress. Today, algorithmic market makers account for over 60% of equity trading volume and provide abundant liquidity in normal conditions, quoting tight spreads and deep order books. But these firms have minimal presence obligations and activate kill switches when volatility spikes, pulling out when liquidity is needed most. Researchers have called this "the liquidity illusion": markets that appear deep and liquid in calm periods but become thin and unreliable under stress, with bid-ask spreads widening and order book depth collapsing.

ETFs amplify this illusion. With over $6.2 trillion in U.S. assets, ETFs account for 31% of equity trading volume and create the appearance of liquid access to every market and sector and asset class. But research has shown liquidity shocks at the ETF level transmit to underlying securities through the arbitrage mechanism, with spillovers increasing during crises. During the August 2015 China selloff, major ETFs traded well below their net asset values. In extreme stress, the arbitrage mechanism breaks down entirely, and ETFs amplify volatility rather than dampen it. A liquidity shock in a single ETF forces price movements in every one of its hundreds of underlying holdings, regardless of those holdings' fundamentals.

Leveraged ETFs add another layer. These products, which reached a record $117 billion in assets in 2024, must rebalance daily to maintain their target leverage ratio. On September 3, 2024, leveraged ETFs sold an estimated $15 billion in equities in a single session, contributing to a 3% plunge in the Nasdaq 100. The rebalancing is concentrated in the final 30 to 60 minutes of trading, creating predictable end-of-day volatility surges that other algorithms front-run. A 2x leveraged ETF tracking an underlying that drops 10% and then recovers 11.1% ends up losing 1.82% even though the underlying is flat. This is volatility decay, and it quietly erodes leveraged positions during choppy markets.

The calm, liquid, low-volatility market that prevailed for much of 2025 was not just good fundamentals. It was partly an artifact of the same leverage, gamma positioning, and algorithmic market-making that would amplify the eventual reversal. Low volatility attracted more capital into volatility-targeting strategies, which increased their equity exposure, which suppressed volatility further, which attracted even more capital. This reflexive loop is the inverse of the crash dynamic. Equally mechanical, equally divorced from fundamentals, and bound to reverse.

The February 2026 selloff was a correction after a long period of low volatility and rising prices. Record margin debt, extreme gamma exposure from 0DTE options, globally interconnected algorithmic trading, and historically high asset valuations created a system primed for exactly this kind of violent repricing. The triggers were almost incidental. Anthropic released a product. A Fed chair was nominated. A jobs report disappointed. None of these individually warranted a multi-trillion-dollar global selloff. But filtered through leverage, algorithmic feedback loops, and cross-asset margin contagion, they triggered a cascade that erased wealth at a pace no rally could match.

A few things are worth taking from this. Market cap destruction is not the same as money moving. The trillions that "vanish" during a selloff are primarily mark-to-market revaluations of unrealized gains, driven by marginal transactions that reprice every outstanding share. When $83 billion flowed into money market funds in a single week, that was real money seeking safety. The other trillions simply ceased to exist.

The modern market's architecture is also inherently asymmetric. Margin calls force selling but nothing forces buying. Volatility-targeting funds sell into declines but ramp up slowly during recoveries. Negative gamma amplifies downward moves while positive gamma merely dampens upward ones. The machinery has more mechanisms pushing prices down quickly than pulling them up.

And diversification fails when you need it most. The yen carry trade connects Japanese monetary policy to Nasdaq momentum stocks. The basis trade connects Treasury futures to hedge fund equity positions. ETF arbitrage connects the VIX to Korean semiconductor stocks. In a crisis, these hidden linkages activate simultaneously, turning a portfolio of "uncorrelated" assets into a single correlated bet on risk appetite.

The billionaires who lost $62 billion in five weeks, the retail investors whose leveraged Nvidia bets went underwater, the systematic funds that mechanically sold into the decline, and the market makers who earned record profits from the volatility are all participants in the same system. That system builds wealth slowly and destroys it quickly. Not because of any conspiracy or market failure, but because of the math of compounding losses, amplified by leverage, accelerated by algorithms, and transmitted globally at the speed of light.