Nasdaq loses more than $4 trillion in a week! After "Black Friday" in global stock markets, the market is beginning to anticipate an oversold rebound in the AI industry chain.
The S&P 500 plunged 3.70% on Friday to 6,910.39 points, closing lower for a seventh consecutive session and, together with the Dow, which fell 1.93%, both retreated more than 10% from their record closing highs, confirming entry into a technical correction. The Nasdaq Composite fell 5.28%, erasing its year-to-date gains; the Nasdaq 100 dropped 5.61% and stood roughly 19.84% below its record closing high on October 6, at one point falling more than 20% intraday to enter what is known as "technical bear market territory."
As of the U.S. stock market close on November 20, Eastern Time, the U.S. stock market and even global stock markets evolved from Thursday's tech stock pullback and the three major U.S. indices all falling more than 2% into a global stock market "Black Friday" dominated by a crash-style plunge in the global AI computing infrastructure theme and broad technology stocks.
Following four extremely weak trading days this week, the S&P 500 fell 3.70% on Friday to 6,910.39 points, closing lower for a seventh consecutive session, and together with the Dow, which fell 1.93%, both retreated more than 10% from their historical closing highs, confirming entry into a technical correction. The Nasdaq Composite fell 5.28%, erasing its year-to-date gains; the Nasdaq 100 fell 5.61% and was about 19.84% below its October 6 historical closing high, at one point falling more than 20% intraday to enter so-called "technical bear market territory." For the Nasdaq Composite, which covers a broader range of technology stocks, the combined market capitalization of Nasdaq Composite constituents shrank by more than $4 trillion during the week ended November 20, Eastern Time.
Two companies that had been expected to turn around AI trading sentiment were sold off at the same time: Anthropic fell to as low as $83.05 on its second trading day, just about 5 cents from its issue price, and closed at $88.91, down 8.50%; Nvidia plunged 14.01% after reporting earnings with revenue doubling. The VIX fear index surged intraday on Friday to 40.03meaning the intraday fear index at one point hit its highest level since the April 2025 tariff turmoiland Goldman Sachs' trading desk fear gauge also soared on Friday to 9 pointsout of a maximum of 10 on that Goldman benchmark fear index.
The core logic of this round of global technology stock selling, including "Black Friday," is that global investors have begun to distinguish between the sales revenue chipmakers have already realized and the investment returns that large AI computing buyers such as AI neoclouds and cloud computing giants have yet to realize. For data center operators, GPU delivery is only the beginning of capital investment; only when clusters are commissioned on schedule and paid utilization and rental rates meet expectations will projects have enough cash flow to cover power costs, repay debt, and recover equipment investment.
Interest rate futures market pricing for a 25-basis-point Fed rate hike in December remains close to 100% certain, and expectations for a 50-basis-point Fed hike in December and further hikes in early next year continue to rise, further raising the return threshold for new projects and compressing the valuation premium investors are willing to pay for long-term profits. In addition, for the direction of global capital flows, as heavily crowded leveraged AI longs sharply cut exposure and trend-following "fast money" quantitative forces (represented by CTA quantitative strategies) mechanically reduced positions on a large scale as prices weakened, doubts about capital returns spread from model companies and chipmakers to the cloud computing leasing and data center power equipment chain, rapidly amplifying the severity of the selloff across the entire AI industry chain.
With the market this week too fearful, especially as leaders in the AI computing industry chain and broad technology stocks fell too sharply, many institutions are anticipating a sudden violent rebound. Many top Wall Street strategists have explicitly turned bullish on an oversold rebound in U.S. and even global technology stocks during Thanksgiving week (November 23 to 27, i.e., next week).
From breaking down on Thursday to stampede-style plunge on Friday, the "AI computing bellwether" Philadelphia Semiconductor Index fell 17% in a week
To gauge the depth of this stock market correction, the U.S. Philadelphia Semiconductor Index, which in recent years has earned the title of "AI computing bellwether," is more representative than the Nasdaq.
The Philadelphia Semiconductor Index plunged 10.73% on Friday to 8,408.59 points, down 42.54% from its June 22 historical closing high. It fell 17.13% for the week, the worst week since September 2001, with constituent companies losing a combined market value of about $2.3 trillion. U.S.-headquartered memory chip leader Micron Technology plunged 13.89% in a single day to $594.51, down a staggering 51.01% from its historical highest closing price set in June, amounting to a halving-style crash.
In the same week, the S&P 500, Dow, and Nasdaq (Nasdaq Composite) fell about 8.6%, 8.0%, and 10.0%, respectively, with the Nasdaq posting its largest weekly decline since March 2020. U.S. stocks traded 25.19 billion shares on Friday, about 29% above the average daily level of the past 20 trading days, and decliners outnumbered advancers by 4.27 times on the Nasdaq market, as selling pressure spread wildly from a handful of AI mega-leaders to the entire stock market, with panic spreading to every corner.
Asian stock markets had already undergone a round of repricing before New York officially opened for trading on Friday. Taiwan's weighted index, which in recent years has fully benefited from the AI infrastructure boom, fell 6.2% that day and 17.5% for the week, while South Korea's KOSPI and Japan's Nikkei 225 fell about 5.2% and 5.1%, respectively, that day. Subsequently, Europe's Stoxx 600 fell 2%, and the MSCI All-Country World Index fell 8.1% for the week, the largest weekly decline since early 2020.
Taiwan's wafer foundry and server manufacturing chain, South Korea's memory chips, and Japan's vast semiconductor equipment chain have different revenue recognition stages, yet all are exposed to the same variablewhether North American cloud computing companies' next batch of hyperscale AI data center construction projects can, after the Democratic Party controls both the Senate and House, secure large-scale financing on schedule, be powered up, and continue to frantically procure core equipment related to AI infrastructure.
As mentioned above, the core of this round of frantic market selling and the brutal "Black Friday" plunge is that investors have begun to distinguish between the revenue data chipmakers have already realized and the investment returns that AI computing buyer forces have yet to realize.
For data center operators, GPU delivery is only the beginning of capital investment; only when clusters are commissioned on schedule and key revenue-generating data from downstream model companies and other AI applications, paid utilization, and rental data meet expectations will large AI infrastructure projects have enough cash flow to cover power costs, repay debt, and recover equipment investment.
The market's doubts about the returns on AI capital expenditure center on whether enterprises' paid demand for AI applications can continue to grow substantially faster than expected.
For model and application companies, if subscription and API revenue growth can continue to outpace the computing costs required to provide services, renewing leases and expanding GPU purchases will have a more solid commercial foundation. For data center operators, only when clusters are commissioned on schedule, customers pay rent as agreed, and existing contracts are renewed at sufficient rental rates or taken over by new customers upon expiration will projects be able to cover power and O&M costs, service debt on schedule, and recover investment within the equipment's economic life.
Nvidia's revenue doubled, so why is the entire AI infrastructure chain from GPUs to backup power plunging?
Nvidia's third-quarter revenue was $115.8 billion, up 103% year over year, above Wall Street's estimate of about $109 billion; data center revenue rose 110% year over year to $107.6 billion, and the company sharply raised its growth outlook for the next fiscal quarter and next fiscal year above expectations. The earnings report confirmed how sought-after the computing power already delivered is, yet the stock price is reassessing how high a return customers can earn to continue expanding.
On Friday, AMD fell about 13.3%, Broadcom fell 12.2%, cloud computing operators CoreWeave and Nebius fell about 12% and 16%, respectively, and even Caterpillar, which sells power generation and backup power equipment, fell about 5%. For chipmakers, the risk is that customers delay their next round of purchases; for operators that have already borrowed to build data centers, the same delay means rental income is pushed back while interest and equipment costs have already begun to accumulate.
Nvidia's stock price also came under pressure independent of the industry cycle. After Jensen Huang's departure, the management transition is still underway. On Friday, Reuters cited two people familiar with the matter as saying that U.S. SEC enforcement staff are examining potential accounting issues at Nvidia; before the report was published, the stock had already fallen more than 6%, and the decline widened after the news emerged.
The specific matters under review have not been made public, and the regional revenue attribution and customer financing arrangements that investors continue to focus on cannot currently be understood as major issues already identified by U.S. regulators. Horizon Investment Services CEO Chuck Carlson had originally hoped that Nvidia's earnings and Anthropic's listing would allow the market to digest pent-up doubts; Friday's close showed that neither of these two positives resolved investors' concerns about the next round of capital returns.
Cetera Chief Investment Officer Gene Goldman believes the two test different questions: the top chip company's earnings have already proven strong current profitability, while Anthropic's listing tests how high a valuation investors are willing to pay for future growth.
When the 10-year U.S. Treasury yield is still near 5% and the market expects the Fed to continue raising rates, even if AI demand remains strong, investors will ask how much of future years' growth has already been priced into the stock.
Ameriprise Financial Chief Market Strategist Anthony Saglimbene noted in a Friday closing report that investors are reassessing the financing capacity of computing power buyers: borrowing costs are high, whether credit funds can continue to be available is in question, and Nvidia's management turnover adds another variablenamely, how much financing support the company will be willing to provide to small and medium-sized customers going forward.
Treasury Partners Chief Investment Officer Richard Saperstein noted that the same slowdown in data center investment does not have the same earnings impact across the industry chain.
Tighter financing for new projects will technology companies' earnings growth expectations, but hyperscale cloud service providers that already own a large number of operating data centers can still serve customers with existing facilities; if new capital expenditure slows, their free cash flow may instead improve.
Treasuries rise, Fed still hawkishwho determines the next "selling storm"?
Bill Northey, senior investment director at U.S. Bank Wealth Management, said selling pressure was still accelerating near Friday's close, with short-term prices more dominated by sentiment, and it will take longer for economic resilience and corporate earnings to show their support for stock prices.
Goldman Sachs' trading desk positioning model further explains why selling pressure may persist even if the index stabilizes: in a scenario where stocks move sideways over the next week, trend-following strategies such as commodity trading advisors (CTAs) may still net sell about $65.8 billion of stocks, and if the decline continues, potential selling could reach $80.6 billion. This means that even if corporate earnings expectations do not deteriorate in tandem, systematic deleveraging triggered by earlier declines may still amplify short-term price volatility.
On Friday, safe-haven funds bought U.S. Treasuries, with the 10-year yield falling about 10.4 basis points to 4.948%, while gold and silver fell 2.81% and 4.22%, respectively, and Bitcoin at one point fell below $60,000. Although Brent crude fell about 4% for the week, that was still not enough to change near-term rate pricing.
Fed Governor Cook believes previous rate hikes were insufficient, while Philadelphia Fed President Paulson advocates prudent progress while remaining open to raising rates at two consecutive meetings. Federal funds futures still price a 25-basis-point December rate hike as nearly certain. The decline in long-term bond yields reflects investors buying safe-haven assets, while the actual financing cost of AI projects also depends on credit spreads. U.S. October PCE data to be released at 21:30 Beijing time on November 25 will test whether the market has reason to lower rate hike expectations after the selloff.
Goldman Sachs' estimates explain why selling may persist even if the market stabilizes. The bank estimates that trend-following strategies have sold about $57 billion of stock positions over the past month, with U.S. equity positioning turning to a net short of $19 billion. Over the next week, if stocks rise, the model still expects trend strategies such as commodity trading advisors (CTAs) to potentially sell $28.6 billion of stock positions; in a sideways scenario, $65.8 billion; and in a scenario of continued declines, the expected buying is about $80.6 billion. Goldman's research team said the core reason is that a downtrend accumulated over multiple weeks will not reverse immediately because of a one-day rebound. Investors should note that $80.6 billion is a quantitative model estimate under the decline scenario, and public disclosures have not limited it entirely to U.S. equity selling.
Five trading observationshow to distinguish an oversold rebound from a true stabilization of AI valuations?
On Friday, Anthropic intraday came within just 5 cents of its $83 issue price, and the Philadelphia Semiconductor Index plunged 10.73%. After such a sharp drop, short covering or trend funds slowing their selling could both bring about a rebound. The following five trading observations are meant to answer whether, if this oversold rebound in the AI industry chain and broad technology stocks occursmost institutional investors are beginning to expect global technology stocks to stage a technical oversold rebound next week after the massive plungethere is enough buying and earnings foundation to support it continuing.
Nathan Peterson, director of derivatives research and strategy at the Schwab Center for Financial Research, explicitly turned bullish on U.S. stocks next week in a trading outlook published Friday. He noted that at the time of writing, the relative strength index (RSI) for both the S&P 500 and the Nasdaq Composite had fallen to 26, while the VIX had risen above 30, indicating short-term selling was nearing extremes and the probability of a technical rebound was higher than usual. He is particularly focused on follow-through after next Monday's open: if the index opens lower but closes higher, or if it remains under pressure on Monday and quickly recovers losses on Tuesday, that would show more clearly than a simple gap-up that buyers are willing to absorb sellers.
TheStreet Pro technical analyst Helene Meisler also points to next week as the rebound window. She said that even if the Nasdaq continues to close slightly lower, the momentum indicators she tracks may turn higher by midweek next week. In her experience, tradable rebounds are more likely when downward momentum fades and investor sentiment has already clearly turned pessimistic. The firm's portfolio manager Chris Versace noted after Friday's close that next Wednesday's U.S. personal consumption expenditures (PCE) inflation data could be a catalyst: if inflation is weaker than expected, market concerns about further Fed rate hikes may ease, providing conditions for a rebound in technology stocks that have already entered oversold territory.
In the short term, as underwriters complete overallotment replenishment, who is willing to take real money to buy near the issue price, and whether systematic selling pressure can ease after volatility subsides, will determine whether the market can first stabilize prices. Further out, only when the Nvidia CEO controversy gradually subsides and the new management lineup and product line outlook gain market recognition, how much revenue increased model calls from enterprise customers can generate, whether gross profit earned per unit of computing power can improve, and whether computing power operators can finance the next batch of GPUs at affordable interest rates will determine whether investors have reason to raise valuations for the AI industry chain again.
Anthropic's $83 reveals a price disagreement between two types of buyers. Investors allocated shares at the issue price were still up 7.12% as of Friday's close; investors who bought at the $111.50 opening price on the first day and held have already lost 20.26% over two trading days. The "buyers' strike" mentioned by Forte Securities refers to some institutions that did not receive allocations and face no index-tracking pressure being willing to wait for lower quotes. Vega IS even explicitly said it would consider buying only near or below the issue price. Demand is still there, but the premium buyers are willing to pay has narrowed sharply.
The record fundraising amount did not leave a market-maker of the same scale in the secondary market. After underwriters fully exercised the 147 million-share greenshoe, the gross transaction amount of the IPO and concurrent private placement totaled $119.5 billion, including proceeds from the sale of existing shares, with the company itself receiving gross proceeds of about $117.6 billion. Underwriters acquired additional shares at the issue price, already covering the overallotment position. If the stock price approaches $83 again, the buying that might previously have come from covering this position through the open market no longer exists, and whether the issue price can hold depends more on genuine quotes from new investors.
Growth in model call volume can support the next round of chip orders only if it translates into sufficient revenue per unit of computing power. Anthropic disclosed that its annualized revenue run rate had reached $81 billion as of the end of October, but the company typically purchases computing power 12 to 18 months in advance, so short-term costs are relatively fixed. If the additional call volume brought by model price cuts is enough to raise utilization, more paid tasks can dilute idle costs; if each task consumes more inference resources, or if the price per call falls faster than unit inference cost, revenue expansion may also come with margin pressure. This is exactly why the market has traced questions from Claude's enterprise demand all the way to GPU and high-bandwidth memory orders.
The credit market has already begun pricing different AI projects differently. A CoreWeave bond issued in June at a 9.625% yield saw its secondary market yield approach 15% by mid-November, directly affecting the pricing reference for new financing and refinancing. Meanwhile, Apollo and Goldman Sachs on Friday still finalized about $44 billion of GPU financing for SpaceX, including a tranche of senior secured notes due 2033 with a yield of 8.259%, with the related lease expected to start when equipment is delivered. The two financings have different collateral and maturities, but the transaction results reveal what capital is still willing to support: when equipment is delivered, who bears the rent, and what recourse creditors have to assets and cash flow all directly affect financing terms.
A VIX at 40 can explain why a rebound may be violent, but cannot alone judge the bottom. The VIX ultimately closed Friday at 35.89, up 26.20%, the highest close since the April 2025 tariff turmoil. Mandy Xu of the Chicago Board Options Exchange estimates that more than 80% of the increase can be explained by the S&P decline and its chain effect on option pricing. Under high volatility, even a slight weakening of selling could bring a sharp rebound. What better tests the quality of a rebound is whether the number of stocks hitting 52-week lows when the index recovers clearly falls from Friday's high, whether credit spreads for computing power operators narrow in tandem, and whether new data centers can be delivered on schedule and begin collecting rent. The three measure breadth of buying, financing capacity, and cash flow realization, respectively.
Related Articles

Ming-Chi Kuo: Google to cut 2025 Broadcom-designed TPU shipments by 20%, plans to increase MediaTek share

Fed Officials Sound Increasingly Hawkish! Governor Cook: Further Rate Hikes Needed Immediately; the Cost of Delaying Action Is Greater

AI short trades begin to take profits? "Big Short" Burry says put options on Nvidia and semiconductor ETF both halved.
Ming-Chi Kuo: Google to cut 2025 Broadcom-designed TPU shipments by 20%, plans to increase MediaTek share

Fed Officials Sound Increasingly Hawkish! Governor Cook: Further Rate Hikes Needed Immediately; the Cost of Delaying Action Is Greater

AI short trades begin to take profits? "Big Short" Burry says put options on Nvidia and semiconductor ETF both halved.






