U.S. stocks, gold, and Bitcoin all fell simultaneously--this is not rotation, but money is disappearing.

CN
4 hours ago

On November 20, U.S. stocks, gold, silver, copper, and cryptocurrencies all fell on the same day.

This is not the kind of "rotation where capital moves from one type of asset to another due to problems." This time, there isn't an asset stepping in to take up the slack.

The Nasdaq 100 saw a single-day drop of 5.6% and a weekly decline of 10.4%, marking its worst weekly performance in over six years, having fallen more than 20% from its intraday high on October 6, officially entering bear market territory. The S&P 500 declined by 3.7% for the week, dropping over 10% from its peak, closing below its 200-day moving average for the first time since April. Technology stocks lost more than $4 trillion in market value over the week.

On the same day, spot gold fell nearly 3%, hitting its lowest level in over a year; silver plummeted even more sharply than gold over the week. Copper and tin also registered declines. Bitcoin dropped below $60,000, reaching its lowest point since September 2024.

The VIX spiked to 40 points, the highest since April 2025.

This is a liquidity contraction, not asset rotation. This distinction is crucial as the responses are completely different.

The context of this article

01

Where did the money go?

02

Why did hedges fail?

03

When will the turning point be?


01 · WHAT ACTUALLY FELL

First, let's accurately state that "everything fell."


Many people see that "everything is falling," but the actual situation is more nuanced: the degree and order of the declines hide the answers.

Consider the weekly ranking:

Technology Stocks

The Nasdaq 100 dropped 10.4%, entering bear market.

Silver

Plummeted over 7% in one week, briefly reaching $41.50.

Gold

Fell nearly 9% in one week, breaching the $3,600 and $3,500 levels.

Bitcoin

Cratered from a high of $62,341 to $55,134, falling below $60,000 in just one week.

Copper

Dropped over 3%, hitting a seven-month low.

Crude Oil

WTI rose 1.48% during the session, the only asset that went against the trend.

Note the last point. Crude oil is the only asset that went against the trend. What does that suggest?

If this were truly a "panic of overall recession," crude oil should have suffered the most—recession means a collapse in demand. But it didn’t drop; instead, it rose. Therefore, the main cause of this round is not the recession expectation on the demand side, but financial conditions tightening.

The rope that is holding everyone down is the U.S. dollar.


02 · WHERE THE MONEY WENT

Where did the money go: the answer is "it didn't go anywhere, it has been reclaimed."


This is the most critical section of the article.

The question "where did the money go" is problematic in itself. It presupposes a premise: that money flowed from asset A to asset B, so finding B would suffice.

However, during a phase of liquidity contraction, the money didn't "transfer," it "disappeared."

This sounds esoteric but is actually a principle of accounting.

Imagine you have a U.S. Treasury bond, using it as collateral in the repo market, borrowing money to buy stocks. Meanwhile, the stocks act as collateral at the brokerage, allowing you to borrow additional funds to purchase Bitcoin. In these three rounds of borrowing, each step creates a "financial asset," inflating your balance sheet wealth threefold, but the underlying asset is just that one U.S. Treasury bond.

Now, if the dollar rises, financing costs increase, or a certain step requires additional margin—either you have to put up more money, or you must liquidate your position.

When you close out your stock position, that "magically created purchasing power" disappears at the same moment. It has not flowed into gold, government bonds, or anywhere else. It simply never existed.

This is why we see "all assets falling simultaneously": the purchasing power behind them is a different projection of the same leverage.

The data supports this picture:

01

Bank of America's Hartnett stated that U.S. households' stock wealth shrank by approximately $3 trillion in the past month, but has increased by about $9 trillion so far this year.

02

The stock allocation ratio of Bank of America's private clients has dropped to 65.2%, the lowest in seven months—wealthy individuals are quietly reducing their holdings.

03

The U.S. dollar index is around 103.8, reaching a new high for the year, which Hartnett refers to as the "destructive force of global liquidity."

04

Emerging market bonds saw an outflow of $1.5 billion in one week.

So where did the money actually go?

There are only three real destinations, none of which involve “making money”:

First, repaying debt. Deleveraging means money returns to creditors and exits the financial system. This is the largest portion.

Second, buying U.S. dollars in cash. This is the most counterintuitive aspect of this round—doing nothing but holding cash.

Third, buying government bonds. On the same day that U.S. stocks fell sharply, capital flowed into U.S. Treasuries. The 30-year yield dropped 10 basis points to 5.30%, the 5-year dropped 15 basis points to 4.70%, and the 2 to 7-year maturities all fell by at least 15 basis points, with the 10-year yield briefly dropping to 4.94%.

Falling yields mean that bond prices are rising—on the day when all assets were declining, U.S. Treasuries were among the few that rose. This is a clear indication of inflows into safe-haven assets.

Moreover, note the direction: safe-haven money is going into bonds, not into gold. Even though both are “safe-haven assets,” this round saw demand for bonds and selling of gold.

Thus, the accurate answer is not “where did the money go,” but: most of the money evaporated, and the small portion left went into U.S. Treasuries and U.S. dollars.

This also explains why gold fell. It is not that gold's status as a hedge has failed; rather, its short-term opportunity cost became too high, and it itself became a source of liquidity to be sold off—one key aspect in reports from Xinhua Finance stated: when U.S. stocks were correcting in late October, some capital exiting the stock market directly flowed into high-yield U.S. Treasuries, and there were even occurrences of “some investors selling gold to raise cash.”

In a liquidity crisis, gold's primary identity is not as a safe-haven asset, but as something that can be sold.


03 · THREE ENGINES

Three engines, but only one main switch


Breaking down the causes reveals three threads.

The first thread: a strengthening dollar draining global liquidity

The U.S. dollar index hit its highest point of the year, making it the most crucial variable in this round.

Why can a strong dollar crush all assets? Because it simultaneously does three things: increases the opportunity cost of holding gold (a non-yielding asset, while the dollar earns interest); elevates the dollar cost of all non-U.S. assets; and significantly raises the debt repayment pressure for emerging markets and companies borrowing in dollars.

Hartnett argues that the dollar's surge combined with gold dropping below $4,000 signals a withdrawal of global liquidity.

The second thread: re-pricing interest rate hike expectations, killing all future cash flows

October CPI rose 0.4% month-on-month, and core CPI rose 0.3%, both exceeding expectations by 0.1 percentage points. Following the data release, bets on a rate hike in December rose sharply.

Then, the minutes from the Fed's October meeting released a more aggressive signal. The interpretation from Gu Fengda, chief analyst at Guotai Junan Futures, is worth quoting:

Almost all committee members believe further rate hikes are necessary, with some explicitly stating that December should be “more than 25 basis points.” More crucially, the Fed has rarely identified AI infrastructure investments as the third largest source of inflation.

This statement is the pivot point of the entire narrative.

In the past two years, AI was a "growth story." Now, the Fed defines it as a "source of inflation."

The same thing has shifted from being an asset to a problem. All valuation models based on AI need to be recalculated.

The third thread: the collapse of the AI narrative itself

This thread is the most direct.

What happened over the week was densely packed: NVIDIA saw its largest single-week drop since October 2008 (-22.92%), evaporating about $2.65 trillion in market value from its peak of nearly $60 trillion in October; after CEO Jensen Huang's resignation on November 12, the stock price fell for seven consecutive days; subsequently, The Wall Street Journal reported that the SEC is investigating NVIDIA, with Reuters further citing insiders who stated that law enforcement agencies are examining its accounting issues (it’s essential to emphasize: an investigation does not imply wrongdoing).

AMD saw consecutive sharp declines following a short-seller report, the Philadelphia Semiconductor Index fell approximately 10.7% in a single day. Micron dropped about 14%, Broadcom fell over 12%, and CoreWeave and Nebius saw declines of approximately 12% and 16%, respectively. Even Caterpillar dropped by 5%—because data centers need generators.

The most significant statement came from Anthropic. This AI star company saw a 17.1% rise on its first day of trading, only to fall back near the issue price the next day, closing down 8.5% and at one point being only 5 cents from breaking even.

Chuck Carlson, CEO of Horizon Investment Services, made an astute comment:

NVIDIA’s reported earnings might be one of the best quarterly performances in corporate earnings history… I hoped that Anthropic's IPO would serve as a clearing event for market sentiment, but it did not play that role.

One of the best performing quarters in history, but the worst in history for stock price. This combination indicates that the market is no longer willing to pay for "good"; it is paying for "how long can it remain good."

One engine, three parts

These three threads appear independent, but they are actually three components of a single switch: U.S. dollar liquidity.

A strong dollar → high financing costs → skepticism about the capital expenditures of AI-dependent companies → overvaluation assets are devalued → asset prices drop → need to meet margin calls → sell anything that can be sold (including gold) → U.S. dollar strengthens further.

This is a positive feedback loop, not three independent events.


04 · WHY SAFE HAVENS FAILED

Why did this round of "hedging" fail?


Many people are confused about one question in this round: Isn’t gold a safe-haven asset? Why did it drop even more sharply than stocks?

The answer has two layers.

In the short term, gold is determined by interest rates, not hedging.

In an environment where rates are rising rapidly and real interest rates are elevated, the opportunity cost of holding a non-yielding asset is extremely high. At this time, gold is treated as “the most easily sellable asset” rather than “the safest asset.”

Silver is even more extreme. Its market size is smaller and also suffers from the industrial demand re-evaluation due to the slowdown of AI data center construction, which amplifies volatility—this also explains why silver has fallen harder than gold.

In the long term, gold is still determined by U.S. dollar credit.

One detail that most people may overlook is: central banks and ETFs are still buying.

The World Gold Council reports that in the third quarter, global central banks and other official institutions net purchased 321 tons of gold, up 42% year-on-year; global physical gold ETFs increased their holdings by 211 tons. In October, while gold prices fell, global gold ETF holdings still increased by 19 tons to 4,275 tons. European and Asian funds recorded net inflows, and China's gold reserves have increased for 24 consecutive months.

Officials are buying, while retail investors are selling. This disparity itself indicates a difference in the time scale.

The real short-term risk does not lie in U.S. dollar credit, but in the physical demand side: in the third quarter, global gold jewelry consumption dropped 17% year-on-year, and demand for gold bars and coins fell 19% year-on-year. Weak physical demand has limited the upside potential; central bank purchases of gold have supported the central tendency. These two forces are hedging against each other, resulting in “falling sharply, but not collapsing.”


05 · WHAT TO DO

Four actions traders should take now


This section will not discuss specific price points, only actions, prioritized.

First: clarify whether your current exposure reflects the same thing.

As mentioned earlier, all assets falling simultaneously is due to their underlying connection by the same leverage.

Thus, do the stocks, Bitcoin, and gold you currently hold reflect the same risk?

If your answer is “I hold tech stocks, crypto, and gold, so I am diversified,” you may be mistaken—in this round, the correlation between these three assets is close to 1. Nominally diversified, but effectively concentrated.

Real effective diversification reduces total leverage, not just spreading it across three assets.

Second: don’t sell what you shouldn’t sell at the worst time.

The most typical mistake during liquidity contraction is “selling the most liquid asset to meet margin calls.” This is the reason for gold's decline in this round—it’s not that the fundamentals have deteriorated, but that it is the easiest to sell.

If you need cash now, consider three questions: what is being sold "because I need money" (forced), and what is being sold "because I am bearish" (active). The former is a stop-loss; the latter is a decision. Confusing forced sells as active decisions is the easiest way to lose money.

Third: treat “cash” as a position.

Many people view cash as “not betting.” At this stage, cash is a position, and it might be the best position.

The reasoning is strong: as the dollar strengthens and financing costs rise, the yield on cash increases while the discount rate for other assets also rises. You earn interest while waiting for others to be forced to sell.

The cash ratios of Bank of America’s private clients and cash levels reported in fund manager surveys have historically been reliable bottom signals. We are not at that position yet.

Fourth: distinguish between “falling” and “buying.”

This is the hardest point, and also the most easily confused one.

A price drop does not equal being cheap. This round of decline is about “the pricing of future cash flows,” not “the company's operations.”

AI demand is still growing, and NVIDIA's performance is indeed the best in history. What has changed is not the business, but how many times the market is willing to pay for that business.

The "multiple willing to pay" depends on interest rates.

Therefore, the true buying signal is not “how much it has dropped,” but “when will interest rates stop rising.” This clue must be closely monitored.


06 · WHEN IT TURNS

When will the turning point be: focus on these five points, not on prices.


I do not predict specific time points, just listing observable conditions. The logic behind these five points is: this round is driven by liquidity contraction, so we need to wait for that to end.

01

The U.S. dollar index hits its peak. This is the main switch. Hartnett projects the dollar peak will be in February or March next year; Goldman Sachs' forex team expects the euro to the dollar to fall to 1.08 within three months, returning to 1.15 by the end of 2027. If the dollar doesn’t peak, nothing else matters.

02

Real interest rates head downward. Real rates are currently elevated around 2.91%, indicating maximum macro suppression. We need to wait for it to drop.

03

Expectations for rate hikes genuinely loosen. The first cracks are already appearing, as discussed below.

04

Oil prices confirm they have peaked. High oil prices mean inflation won’t decline; if inflation doesn’t decline, the Fed can’t pivot. Hartnett’s first entry condition is WTI breaking below $70; now WTI is around $83.4, and Brent is near $90, still about 16% off $70.

05

Can the cash flows from AI capital expenditures materialize? Bloomberg Intelligence indicates that the sales growth rate of AI firms in the S&P 500 is expected to drop from 35% in the third quarter to below 25% by the end of next year, and the ratio of cash flow to capital expenditures has fallen to about one-fifth of what it was three years ago.

The fifth point needs to be elaborated because it is the most easily underestimated.

That set of numbers from Bloomberg Intelligence is more solid than any sentiment: the money these companies earned is increasingly lagging behind their spending.

If the financial reports over the next two quarters can prove “spending starts to turn into profits,” then the AI narrative can be rebuilt. If not, even if liquidity recovers, the valuation of the AI sector won’t return to previous levels—because the assumptions supporting that valuation have been overturned.

The first crack that has already appeared

On November 20, while U.S. stocks plummeted, three signals emerged simultaneously:

U.S. Treasury yields fell across the board, with the 10-year yield dropping to 4.94%; financial media reported that OIS contracts no longer fully reflected the expectation of a 25 basis point rate hike from the Fed in December; the same report stated that traders are no longer fully digesting the Fed's rate hike expectations for December.

This marks the first sign of “the market beginning to doubt rate hikes” since this round of declines.

However, it is important to emphasize: this is merely a sign, not a turning point. The Fed officials’ rhetoric is still hawkish—Governor Cook has explicitly stated that “immediate further rate hikes are necessary; delayed action would be more costly.”

When the policy rhetoric remains hawkish, but market pricing begins to shift dovish, it is a typical characteristic of a “bottoming region,” not a “bottom.” The two are separated by more than a quarter's time.

A historical reference: December 2018

Goldman Sachs traders are bringing up that script again: in 2018, as the Fed continued to raise rates, U.S. stocks fell 20% from their September peak, forcing the Fed to pivot in January.

The script may replay, but there are two key differences:

First, 2018 was the “end of rate hikes,” while this time is “the restart of rate hikes.” The Fed resumed rate hikes in September after a three-year pause, and it has only raised them twice so far.

Second, 2018 did not have the AI variable. This time the Fed has identified AI infrastructure as a source of inflation, which means that even if the economy slows down, the Fed may not pivot quickly—as the source of inflation has shifted from “cyclical” to “structural.”

Therefore, I believe the turning point will be slower than in 2018. This is an evidence-based judgment, not a guess.


∞ · POSTSCRIPT

My perspective


First, this is not the beginning of a bear market, but the settlement of a bull market's tail end. The past two years' trend of “anything bought goes up” was built on cheap dollars. With the dollar expensive now, that logic has expired. The liquidity has retreated, and what retreats first are the assets most reliant on it.

Second, “hedging failure” is not about loss of effectiveness; it is that the definition of hedging has changed. In a liquidity crisis, what can be sold is considered an asset; what cannot be sold is merely paper wealth. Gold has fallen because it is the easiest to sell—not because it is bad.

Third, the worst part of this round may be over, but the best is certainly not here yet. The combination of “loosening rate hike expectations + falling U.S. Treasury yields” on November 20th is the closest signal to a turning point since this decline began. However, the policy rhetoric remains hawkish, and the dollar is still at a high level. The signal has appeared, but the conditions are not met.

Fourth, at this stage, what matters is not who has the best insight, but who is still standing. The duration of the liquidity contraction period depends on the Fed. And when the Fed pivots depends on inflation and employment. Both of these are beyond your control. What you can control is: your own leverage, your own cash, and whether you can wait for that day.


07 · RISKS

Risk disclaimer


This article provides market analysis, does not constitute investment advice, nor an offer or recommendation, and does not promise any returns.

The market data mentioned refers to intraday data as of November 21, 2026, any analysis cannot predict prices, and past performance does not represent future results. The institutional views and data quoted in the article are sourced from public reports and do not represent the author's stance.

Market risks are extremely high, and leveraged trading could result in complete loss of principal. Please make independent judgments and bear all decision-making consequences yourself.

END

I am Zen Wuce, focusing on blockchain investment research and sharing trading insights.

If you found this article rewarding today, feel free to like and share, and let’s refine our trading cognition together!

免责声明:本文章仅代表作者个人观点,不代表本平台的立场和观点。本文章仅供信息分享,不构成对任何人的任何投资建议。用户与作者之间的任何争议,与本平台无关。如网页中刊载的文章或图片涉及侵权,请提供相关的权利证明和身份证明发送邮件到[email protected],本平台相关工作人员将会进行核查。

Share To
APP

X

Telegram

Facebook

Reddit

CopyLink