On Thursday, October 22, risk-off sentiment triggered by escalating tensions in the Middle East swept through U.S. equities, with the AI supply chain suffering the heaviest losses. Investors reassessed the return prospects of massive capital expenditures in AI, leading to selling pressure on Meta, which had recently increased its AI investments. Retail investors had concentrated their buying in AI-related stocks the previous week, making popular tech stocks more vulnerable to sharp declines as risk appetite plummeted. Meanwhile, safe-haven funds flowed into U.S. Treasuries, pushing the 10-year Treasury yield below 5%, indicating that market stress had spread from AI-themed stocks to the broader macro interest rate environment.
Chip stocks became a major casualty. The Philadelphia Semiconductor Index (SOX), which had rebounded about 2% on Wednesday, closed down nearly 7% on Thursday, as did the semiconductor industry ETF SOXX. Among individual stocks, NVIDIA (NVDA) fell 4.4%, AMD (AMD) dropped 8.1%, Lam Research (LRCX) declined approximately 11.8%, and KLA Corporation (KLAC) slid nearly 9.6%. Memory chips, optical communications, and AI cloud computing sectors all suffered widespread declines.
Meta (META) fell more than 8% intraday, as the company’s five-year credit default swap (CDS) spread, a key credit risk indicator, rose to a record high, highlighting investor concerns over the scale of AI investments, future funding needs, and the return timeline.
At the core of this sell-off is the market's reassessment of the costs, financing pressures, and return timelines of the AI investment boom. Although Meta reported third-quarter revenue and profits that exceeded expectations, it further raised its full-year capital expenditure guidance and disclosed that its future spending commitments have risen to $825 billion. Meanwhile, Lam Research sparked concerns among investors about the pace at which semiconductor equipment companies can deliver on their growth projections due to supply constraints.
Meanwhile, attacks on Saudi Arabia pushed Brent crude above $110 per barrel, reigniting concerns over energy costs and inflation risks. Uncertainty around AI returns, supply constraints across the supply chain, and geopolitical risks have prompted investors to reduce exposure to high-valued tech assets and shift toward relatively safer bonds.
Bloomberg noted that Thursday’s market performance was not merely a reversal of AI-driven trading logic; the more prominent feature was a risk-off sell-off driven by heightened避险 sentiment. During the session, the Nasdaq 100 index briefly fell 3%, while the S&P 500 dropped over 1%. In contrast, the equal-weight S&P 500 index rose 0.2%, indicating that selling pressure was concentrated primarily on large-cap growth stocks such as tech giants. Meanwhile, rising bond prices and declining yields further reflected investors’ shift toward safe-haven assets.
Joseph Saluzzi, Partner and Co-Head of Equity Trading at Themis Trading, said, “It feels like a risk-off day, and tech stocks are always the first to be hit.” He also noted that the rise in bond prices confirms the market is adopting a risk-off stance. Investors are concerned that “after the midterm elections, the conflict with Iran could escalate and last longer than expected.”
Middle East risks push oil prices higher, while safe-haven trades hit tech giants.
Attacks on Saudi Arabia have intensified market concerns about an escalation of conflict in the Middle East, pushing Brent crude above $110 per barrel during trading. Rising oil prices have reignited inflation worries, and with the U.S. midterm elections less than two weeks away, higher energy prices may further fuel market concerns about inflation and policy prospects.
Under this backdrop, capital began reducing exposure to risk assets and shifted toward relatively safer government bonds. Tech stocks, which have higher valuations and greater reliance on future growth expectations, came under pressure first. The simultaneous rise in U.S. Treasury prices and decline in U.S. equities on Thursday further highlighted the risk-off sentiment of the day. Therefore, although the sharp decline in the semiconductor index occurred within the AI supply chain, it cannot be simply interpreted as a broad market bearish outlook on AI.
Rising oil prices not only increase cost pressures on the global economy but may also elevate inflation expectations, raising concerns that central banks like the Federal Reserve may have limited room to cut interest rates. Meanwhile, the construction of AI infrastructure itself requires substantial capital, electricity, and equipment; rising energy costs and financing pressures could further impact projections of investment returns.
For AI-related stocks, this creates dual pressure: on one hand, the market demands that companies prove their substantial investments will deliver returns; on the other hand, rising oil prices and potential interest rate pressures may increase future investment costs and compress valuation potential.
The broad decline in chip stocks on Thursday indicates that AI trading has entered a phase where profitability realization, capital efficiency, and financing sustainability are prioritized. Strong demand remains a key pillar for the industry, but demand growth alone is no longer sufficient to sustain valuation premiums for all companies across the supply chain. Over the coming weeks, earnings reports from major tech firms, their capital expenditure plans, and outlooks from chip equipment manufacturers on supply chains and orders will be critical in determining whether AI trading can stabilize.
Chip leaders generally declined, with equipment stocks leading the downturn across the supply chain.
The semiconductor sector sharply declined on Thursday, with chip design, manufacturing, equipment, and memory stocks nearly all under pressure. By close, Broadcom (AVGO) fell 4.3%, Marvell Technology (MRVL) dropped 8.7%, Intel (INTC) slid 8%, and TSMC (TSM) declined 4.5%.
Semiconductor equipment stocks saw more pronounced declines. In addition to Lam Research and KLA, which each closed down nearly 10%, Applied Materials (AMAT) fell 8.2%. Among memory chip stocks, SanDisk (SNDK) dropped 8.5%, SK Hynix (SKHY) fell 8.1%, Seagate Technology (STX) declined 7.9%, Western Digital (WDC) slid 6.4%, and Micron Technology (MU) dropped 6.3%.
Stocks related to optical communications and AI infrastructure also faced selling pressure. By close, Coherent (COHR) fell 12.5%, Lumentum (LITE) dropped 11.5%, Credo (CRDO) declined 10.7%, and Corning (GLW) fell 8.1%. AI cloud service provider CoreWeave (CRWV) closed down approximately 10%, while Nebius (NBIS) fell 7.2%. By close, Dell (DELL) dropped about 6.8%, Super Micro Computer (SMCI) fell 5.2%, and Oracle (ORCL) declined approximately 2.8%.
The broad-based decline across the supply chain indicates that market concerns have expanded from the valuations of individual chip companies to the entire AI infrastructure investment chain. Growth expectations for chip designers, equipment suppliers, memory manufacturers, optical communications companies, and cloud service providers are all closely tied to the ability of major tech firms to sustain investment in AI infrastructure. Once investors begin to question the pace of return on these expenditures, related assets may simultaneously face downward pressure.
Meta increases its capital expenditures; $825 billion in future spending commitments raise financing concerns.
Meta became another focal point of pressure in AI-related trading on Thursday. The company reported third-quarter revenue of $64.6 billion, a 26% year-over-year increase, surpassing market expectations; earnings per share (EPS) came in at $3.83, also above analysts’ forecast of $3.57. However, the strong current results failed to alleviate market concerns over future investment costs, with Meta’s stock dropping more than 8% at midday and closing down approximately 7%.
Meta has raised its 2026 capital expenditure guidance to $135 billion to $145 billion, increasing the lower end of its previous guidance range from $130 billion. The company disclosed that its future spending commitments, including leases, compute trading, and energy agreements, have risen to $825 billion.
Market focus is shifting from whether AI initiatives can grow to how much capital is needed and when they will generate sufficient returns. Meta CEO Zuckerberg stated that the company’s investments are yielding results, its core business remains strong, and products like the AI assistant Muse show growth potential. However, the company also acknowledged that the return cycles for some new AI investments will be longer than those of its core business.
Meta’s credit risk indicators also deteriorated. According to Bloomberg data, Meta’s five-year CDS price rose intraday by approximately 8.9 basis points to around 119.3 basis points, hitting a record high. The rise in CDS prices indicates higher hedging costs for investors against default risk on related debt, reflecting increased market caution regarding Meta’s expanding AI investments and potential financing needs.
Meta also announced a distribution of a new class of non-voting shares to existing shareholders, which may be used in the future for fundraising, employee compensation, and potential acquisitions. This arrangement has further intensified market focus on how the company plans to finance its AI expansion.
Next, Alphabet, Microsoft, and Amazon will陆续 release their quarterly earnings. These major tech companies’ financial results will not only test whether demand for AI remains strong, but also reveal whether capital expenditures continue to accelerate and whether enterprises can increasingly demonstrate that their investments are translating into revenue, profits, and cash flow. For the market, the key issue in AI trading is no longer just how fast demand is growing, but whether the scale of spending is sustainable.
Lam Research faces supply constraints, prompting a reassessment of growth expectations for equipment stocks.
Lam Research became one of the leading decliners in the semiconductor equipment sector. After announcing its quarterly earnings, the stock dropped more than 10% in early trading and closed nearly 12% lower. The company’s guidance for the next period and concerns over supply constraints have made investors more cautious about industry growth expectations.
According to reports, after the earnings announcement, Lam Research's management stated that, absent recent supply constraints, the company’s guidance would have been higher. These supply constraints mean that, even with strong demand for AI-related chips, equipment manufacturers may not be able to deliver products or recognize revenue at the expected pace. This has prompted the market to reassess whether robust orders can be timely converted into actual performance.
The valuation of semiconductor equipment companies heavily depends on wafer fab expansion and investment in advanced processes. Previously, the construction of AI data centers drove demand growth for chips, prompting wafer fabs to increase capacity and supporting order growth and stock price appreciation for equipment providers. However, delays in equipment delivery or adjustments to customers’ expansion plans could impact the timing of revenue realization over the next several quarters.
Peer stocks such as Applied Materials and KLA also plunged sharply, indicating that investors are not limiting their concerns to Lam Research alone. The equipment sector previously benefited from expectations of AI-related capital expenditures, making its stock prices highly sensitive to future growth prospects; when the market begins to worry about supply bottlenecks, the pace of capital spending, and valuation levels, these stocks are prone to amplified corrections.
However, supply constraints do not mean demand for AI chips has weakened. More important now is whether device manufacturers can alleviate delivery bottlenecks and whether large tech companies’ data center investments can continue to support foundry expansions. Upcoming earnings reports and management guidance will help the market determine whether this correction is merely a valuation and positioning adjustment, or whether growth expectations for the AI supply chain are beginning to change.
Retail investors rushed to buy AI stocks in the week before selling off, putting greater pressure on popular trades.
According to Bloomberg, citing data from JPMorgan strategist Arun Jain, in the week leading up to Thursday’s sharp decline—last Thursday through this past Wednesday—individual investors significantly increased their purchases of semiconductor companies, “neocloud” firms offering next-generation cloud services, and storage device companies, while reducing their buying of the U.S. tech “Magnificent Seven.” Jain noted in a report to clients on Wednesday that retail buying volume for individual tech stocks reached its highest level since the market momentum trades began to unravel in July.
Demand from retail investors for tech ETFs has become more pronounced. The net inflow into related ETFs reached the second-highest level on record for a single week, partly due to investors buying the dip in the latter half of last week. On Friday, tech ETFs recorded their largest single-day inflow ever, pushing the week’s total inflow to 2.1 standard deviations above the historical average. VanEck’s Semiconductor ETF (SMH) was also one of the funds attracting significant capital.
The most heavily bought individual stocks by retail investors include SanDisk, TSMC, and Nebius, all of which fell at least 4% on Thursday. Nebius had previously gained attention on social media investment forums; SpaceX stock was also among retail investors’ most favored picks before Tuesday, but after the company released positive earnings and its shares rose 7.8% on Tuesday, retail investors shifted to selling on Wednesday.
These fund flows indicate that retail investors have not entirely abandoned AI trading, but are instead shifting from the "Magnificent Seven" to a broader AI ecosystem. Joe Gilbert, portfolio manager at Integrity Asset Management, noted that as AI ecosystem companies have demonstrated more durable profitability than initially expected, investors have moved toward a wider range of tech stocks. He believes that, in a high-interest-rate environment, tech companies outside the Magnificent Seven could also achieve strong profit growth while maintaining relatively lower valuation multiples.
However, this rotation has introduced new vulnerabilities: when capital floods into popular sectors such as chips, storage, and new cloud services, a sudden escalation in geopolitical risks or a market reassessment of AI capital expenditure returns could trigger synchronized unwinding of previously accumulated positions. While retail investors’ concentrated buying alone cannot fully explain Thursday’s decline, it means that AI-related stocks, previously favored, are now more susceptible to selling pressure when risk appetite reverses.
Safe-haven funds flowed into U.S. Treasuries, causing the 10-year yield to rise above 5.0% for the first time in a month.
The risk-off trading on Thursday was not limited to the stock market. Amid sharp declines in AI-related stocks, pressure on French bond markets, and renewed tensions in the Middle East, U.S. Treasuries strengthened across the board, posting their best performance in over a year. Yields on some maturities fell by as much as 20 basis points, with the 10-year U.S. Treasury yield dropping below the key 5.0% level for the first time in nearly a month. Buying pressure in U.S. Treasury futures further propelled the market higher.
More notably, market expectations for Fed rate hikes have also been pushed back. Bloomberg reported that traders now fully price in a 25-basis-point hike for January next year, delaying it from December. Previously, rising energy prices were typically seen as driving inflationary pressures and higher bond yields; however, on Thursday, investors began viewing high oil prices more as a risk to economic growth and hedged by buying U.S. Treasuries.
David Robin, Interest Rate Strategist at TJM Institutional Services, said: “The situation is changing. Risk-off is one thing; growth risk is entirely another.” This distinction highlights the deeper meaning behind that day’s bond market movement: investors are not only seeking safe-haven assets due to rising risks, but are also beginning to worry that the sustainability of AI investments and the impact of energy price shocks could affect the broader economic outlook.