The AI computing power supply chain staged a powerful rebound as of Wednesday's US stock market close, buoyed by new operating data from new cloud leader CoreWeave — particularly the company's better-than-expected earnings and customer prepayment outlook, which alleviated market concerns about financing for computing power expansion. Additionally, Anthropic's latest prospectus and IPO roadshow revealed explosive revenue expectations on the AI application side and over $900 billion in procurement and leasing commitments tied to large-scale AI computing infrastructure. However, stock market positioning and risk capital sentiment cooling, expected expansion of AI computing power demand, and market confirmation of a bottom remain three distinct signals at different levels, and they have not yet re-established a positive bullish correlation with each other.
At the close on November 11 US Eastern Time, the Nasdaq 100 Index rose 2.28%, while the Philadelphia Semiconductor Index, known as the "global AI computing power trading bellwether," surged approximately 5.7%, with all constituent stocks closing higher. Micron Technology (NASDAQ: MU) and SanDisk jumped 11.89% and 13.69% respectively, NVIDIA (NASDAQ: NVDA) rose about 2.9%, and Broadcom (NASDAQ: AVGO) gained about 4%. In the optical interconnect space, Coherent and Lumentum rose 9.25% and 10.67% respectively. This rally reflects retail and institutional investors focused on "buy the dip" strategies re-entering core AI computing infrastructure segments such as AI chips, storage, and data center high-speed optical interconnects.
However, as the crowding level of global tech stock holdings dominated by AI computing power and semiconductor sectors has significantly declined, these positive fundamental developments have provided support for some degree of buying回流. Yet position details, options pricing, and technical trends have not yet provided consistent signals of a bottom for tech stocks. Behind the strong counterattack of AI computing power leaders: rising expectations for AI computing power demand and easing concerns about AI financing costs. The Philadelphia Semiconductor Index rose about 5.7% and the Nasdaq 100 Index rose 2.28%, jointly highlighting a strong rebound in the AI computing power supply chain, and the logic behind this strong rebound shows that bullish sentiment in the AI computing power supply chain has gained support from operating data.
CoreWeave's earnings exceeded expectations, with the company expecting to receive approximately $9 billion in customer prepayments and deposits over the next three months, alleviating market concerns about financing for computing power expansion. Anthropic's third-quarter revenue grew rapidly, providing validation for paid demand for AI applications, and the approximately $900.9 billion in multi-year procurement and leasing commitments disclosed in its prospectus increased the visibility of long-term computing power demand. The crowding level of tech stock holdings has declined, with implied correlation rising only modestly, indicating that market expectations for a spillover of the selloff to other sectors remain limited. However, the skew indicator measuring relative demand for put options remains in the middle of its range over the past 13 years, and QQQ, an ETF tracking the Nasdaq 100 Index, had not yet broken through its downtrend since late October at that time. Whether the rebound can evolve into a trend reversal depends in the short term on whether the index can break through the downtrend and whether buying can expand to more sectors, and in the medium term on whether computing power orders can be delivered efficiently on schedule and converted into cash flow.
Anthropic's incremental information provides revenue validation from the application side for computing power demand. The prospectus shows that the company's third-quarter revenue reached $16.956 billion, a year-over-year increase of approximately 1,127% and a quarter-over-quarter increase of 46.7%. Its annualized revenue run rate reached $81 billion by the end of October, an increase of $10 billion from the end of September. However, this run rate reflects recent business scale, not full-year revenue guidance. There are also longer-term growth forecasts from Anthropic's core underwriters: according to reports, Goldman Sachs models project Anthropic's revenue to reach $52 billion and $136 billion in 2026 and 2027 respectively, while also expecting free cash flow net outflow of $65.7 billion in 2027, indicating that commercial expansion and infrastructure investment will continue to accelerate in tandem. Anthropic's recently disclosed $900.9 billion in procurement and leasing commitments are mostly related to computing power and infrastructure, of which approximately $451.6 billion is expected to occur before the end of 2030, $70.7 billion will mature within 12 months after the end of September 2026, and the remainder is mainly distributed between 2031 and 2045. It is understood that its Anthropic Direct model separates data center capacity from chip leasing arrangements, first locking in facilities with longer construction cycles, then determining larger-scale accelerator investment based on demand.
For the global AI computing power supply chain, this is undoubtedly a high-intensity positive catalyst. These mean that AI computing power demand has long-term continuity, while revenue recognition still depends on specific delivery节奏: GPUs and ASICs handle computation, HBM provides high-bandwidth data supply, optical interconnects support cluster communication, and power supply and cooling determine whether equipment can actually be put into operation. Long-term procurement commitments support demand expectations, and on-time delivery and sustained utilization rates determine when these expectations translate into supplier profits. The more direct market catalyst on Wednesday came from CoreWeave. The company's third-quarter revenue was $3.661 billion, a year-over-year increase of 168%, and it raised its full-year revenue guidance to $13 billion to $13.3 billion, while expecting to receive approximately $9 billion in customer prepayments and deposits over the next three months, driving its stock price up about 24.7%. This information simultaneously addressed two concerns: "whether demand can be fulfilled" and "where expansion funding will come from." A team led by Bernstein senior analyst Madison Rezaei believes the company's performance was strong and responded to financing concerns. The series of payment dynamics arranged by new cloud-type AI computing power leasing/selling companies is becoming an important link connecting upstream hardware demand with capital market confidence.
Crowded trades have cooled, but key indicators have not yet jointly confirmed a bottom. Deutsche Bank's proprietary statistics show that tech stock positioning has fallen below the lows during the April 2025 "Liberation Day" shock. The institution's senior strategist Parag Thatte mainly attributed the recent decline to forced deleveraging and maintained an overweight stance on tech stocks. At the same time, market expectations for future synchronized stock volatility rose only modestly. Asym Research founder Rocky Fishman believes this indicates that investors still mainly understand the selloff as a sector-level adjustment and have not widely priced in an economic pessimistic downturn affecting most industries. The options market shows that speculative heat has faded but has not yet reached extreme pessimism. Retail activity in buying call options on AI stocks has decreased significantly, changing the trading dynamics that previously supported stock prices. In the scenario where market makers take the sell side of call options and hedge by buying stocks, the ebb in call demand and falling stock prices may prompt them to reduce stock hedges, amplifying selling pressure. In addition, it is worth noting that the one-month standardized put-call skew indicator has only returned to the middle of its range over the past 13 years — highlighting that short-term crowding has decreased, which is conducive to market rebalancing, but is insufficient to judge that the market is already at a bottom-fishing point based on positioning alone. In terms of key technical indicators, QQQ has not yet broken through its downtrend since late October, nor has a DeMark trend exhaustion signal appeared. Wednesday's rise further improved price performance, but internal market divergence remains: total market volume was 14.74 billion shares, below the average of 18.68 billion shares over the past 20 trading days; 159 stocks on the Nasdaq market hit 52-week lows, more than the 77 that hit new highs. Therefore, this rebound has gained support from the AI computing power demand growth outlook and easing financing concerns, and its sustainability will depend on whether the rally can expand from core AI heavyweight stocks to more stocks, and whether the business improvement progress of AI computing power supply chain leaders, cloud computing giants, and software leaders focused on AI application software can continue to offset valuation pressure from high interest rates/long-term US Treasury yields.
After the US market experienced a tech stock crash, various signs indicate that it is still too early to assert that the worst is completely over. The tech-heavy Nasdaq 100 Index hit a record high in early October, and the sharp decline in the following weeks left tech and AI-related stocks about 10% below their recent highs, though stabilization signs have recently emerged. Tech stock positioning, stock correlation, options demand, and technical charts — these four indicators provide a watchlist for judging whether the market is once again near a turning point. Analysts say two of them are encouraging, one is mixed, and the technical picture shows that vigilance cannot be relaxed yet. "The market fell 5% in a single day and closed at the lows, and you cannot then assume the bottom has appeared," said Mark Newton, head of technical strategy at Fundstrat Global Advisors, in an interview with media. "I think this will become a good buying opportunity. But that is not a reason to conclude that the market has officially bottomed," he said.
The overheating phenomena tracked by these indicators are not new. Investors' high concentration in AI trades, continued betting on gains through single-stock options, and overly extended momentum indicators have been prominent market features for months. In early October, as investors began to worry about AI infrastructure spending and its expected returns, the multi-year AI stock rally stalled, and some of these indicators began to pull back. Subsequently, hyperscale cloud service providers building data centers released third-quarter results, revealing signs of project delays and slowing capital expenditure growth, prompting investors to sell shares and bonds of NVIDIA-backed AI infrastructure company CoreWeave. The latest catalyst was the US midterm elections. Strong Democratic performance rekindled concerns about slowing growth, dealing a heavy blow to AI stocks. In recent trading sessions, stock prices have stabilized somewhat, recovering part of the post-election decline. Before the latest pullback, some investors had already begun to worry about the market's reliance on tech and AI stocks. Mega-cap growth and tech stocks have long led the bull market, and investors' high allocation to these stocks has been a topic of market attention for months. Some believe that current positioning conditions may mean the market is poised for stabilization. Deutsche Bank data shows that since late September, investors have been reducing exposure to these stocks while rotating into other sectors. The bank said last week's selloff pushed tech stock positioning below the lows during the April 2025 "Liberation Day" period. "We believe this recent selloff was driven by forced liquidation," said Deutsche Bank strategist Parag Thatte. "Positioning has just fallen below the lower edge of its long-term volatility range, and tech stocks' relative performance has also just fallen below the lower edge of their long-term trend channel. Previously, tech stocks rebounded from these levels," Thatte said, noting that the bank is currently overweight the tech sector. Anthropic's initial public offering in the coming days will test Wall Street's confidence in the AI boom.
US stock market internal trends are not synchronized. Another factor encouraging bulls is that investors' expectations for stock market correlation rose relatively modestly. Correlation refers to the tendency of stocks to move together rather than independently. The AI boom has changed the relationship between stocks. Barclays strategists said in an October 27 report that the three-month realized correlation of the 50 largest stocks in the S&P 500 fell to a historic low in late September. They attributed this to strong stock-specific themes, "especially AI," working alongside macro drivers such as interest rates, inflation, and oil prices. Last week's sharp selloff raised expectations for how synchronized stock movements will be in the coming months, but the increase was limited. "The fact that implied correlation did not jump more significantly suggests investors see this more as a sector-specific selloff rather than a significant economic downturn affecting most or all sectors," said Asym Research founder Rocky Fishman.
Options trading reverses. The sharp growth in demand for call options has been an important feature of this stock market rally. Call options give holders the right to buy a stock at a fixed price, while put options give holders the right to sell. The signal from this market is mixed. Carrick Lane data shows that retail investors bought a record number of single-stock call options this year. This helped support stock prices because market makers — usually large banks and trading firms taking the other side of these trades — buy stocks to hedge their exposure. John Marshall, a partner at Carrick Lane, which manages actively traded options strategies, said this mechanism reversed last week, exacerbating the selloff. "Last week, retail activity in buying call options on AI stocks fell sharply. Our data shows this was the most notable positioning change behind the selloff during election week," Marshall said. He said this change brought the one-month standardized put-call skew indicator back near the middle of its range over the past 13 years. The indicator measures relative demand for put options, so it presents a mixed picture. "The good news: the AI trade no longer looks crowded in the short term. The bad news: the degree of positioning clearing reflected by this indicator is not enough to support buying stocks based on positioning alone," Marshall said.
Technical picture still shows cautious wait-and-see stance. Although the tech stock rebound helped them recover part of the losses caused by the post-election selloff, technical strategists remain cautious. "We have not really seen evidence that the market has been sufficiently cleared out," said Newton of Fundstrat. Newton said the rebound in Invesco QQQ ETF, which tracks the Nasdaq 100 Index, stalled slightly below its downtrend line since late October. Newton said that given short-term trends and momentum remain bearish and no DeMark trend exhaustion signal has appeared — a technical signal indicating a trend has run its course — US stocks are likely to continue fluctuating for about two more weeks. "It is still too early to have confidence that last week's decline has ended," Newton said.
Comments