Can consumer resilience prop up the U.S. stock market rebound? U.S. retail data beat expectations, leading the market to stand firm against the most intense tech-stock selloff in 20 years.

date
22:20 17/11/2026
avatar
GMT Eight
US goods consumption entered the fourth quarter on a firmer footing than the market's previously optimistic expectations, highlighting the strong resilience of US economic growth that relies on consumer spending expansion and reinforcing the "soft landing" expectation for the US economy that Federal Reserve officials have been focused on.
The newly released U.S. October retail sales figures provided strong support for the resilience of U.S. economic growth driven by the AI investment boom, but did not significantly alleviate, or to some extent remove, the increasingly escalating heavy pressure from Federal Reserve interest rates faced by a broad range of tech stocks, including AI computing power and semiconductors. This is also why the strong resilience of American consumers can provide fundamental support for U.S. tech stocks after consecutive adjustments, but is unlikely to drive tech stocks dominated by AI computing power and semiconductor heavyweights toward a new round of valuation expansion. U.S. consumption data and, in a broad sense, the resilience of macroeconomic growth are locked in a tug-of-war with the dramatic allocation adjustment in tech stocks: a Bank of America survey shows that the net overweight ratio for tech stocks fell from 37% to 12%, the largest single-month decline since 2006, but only 3% of respondents expected a recession, indicating that funds are mainly reassessing AI capital expenditure and investment returns. Solid consumption raises the possibility of earnings stabilization, but has not simultaneously lowered the discount rate for tech stocks; after the retail data release, Treasury yields moved higher and the Nasdaq ETF gave back part of its pre-market gains, also showing that interest rates remain an important constraint on the rebound. Consumer resilience can create optimistic macroeconomic tailwinds for the repair of the U.S. tech sector, while a sustained rebound trajectory requires companies to deliver growth through profits and cash flowcompanies that can capture consumer demand and maintain stable profit margins over the long term, as well as AI suppliers with more solid order delivery and funding guarantees, are better positioned to turn the market rebound into a fundamental re-rating. Retail and food services sales released on November 17 reached $748.2 billion, up 6.8% year-over-year and 0.2% month-over-month. Although the headline growth rate was lower than September's 0.8%, the performance excluding autos was stronger, and the control group sales and upward revision to the prior value are especially favorable for fourth-quarter consumption growth expectations. U.S. October retail and food services sales rose 0.2% month-over-month, beating expectations and slightly above the 0.1% median economist estimate compiled by institutional surveys; the previously released September retail sales month-over-month increase was maintained at 0.8%, unrevised. Although headline growth slowed, consumption was more resilient after excluding some volatile items. Specifically, retail sales excluding autos and parts rose 0.6% month-over-month, above the 0.4% expected, while the previously released September comparable increase was revised up from 0.9% to 1.1%. Sales excluding both autos and gasoline rose 0.5%, above the roughly 0.3%0.4% expected in various institutional surveys; the September comparable increase was also revised up from 0.2% to 0.5%, indicating that underlying demand was stronger than the initial reading reflected. 0.2% masks consumer resilience: the real surprise is in the control group and prior revisions More noteworthy for investors is that retail "control group" saleswhich provide an important basis for estimating goods consumption in GDProse 0.6% month-over-month, exceeding the 0.4% expected, and the September increase was sharply revised up from 0.1% to 0.5%. This indicator excludes food services, auto dealers, building materials and gasoline station sales. The stronger-than-expected performance for the month combined with upward revisions to the prior month shows that the slowdown in headline retail sales alone is not enough to conclude that U.S. consumption has clearly weakened. U.S. goods consumption entered the fourth quarter on a firmer foundation than previously optimistic market expectations, highlighting the strong resilience of U.S. economic growth that depends on consumption expansion and reinforcing the "soft landing" expectation for the U.S. economy that Fed officials have been focused on. Official data and media surveys show that the main upside surprise came from core components. The significance of the control group is that it provides an important input for estimating goods consumption in GDP. This indicator excludes autos, gasoline, building materials and food services; it rose 0.6% in October, and combined with the September increase revised up from 0.1% to 0.5%, it means consumption entered the fourth quarter on a firmer foundation than previously estimated. The decline in auto sales can explain part of the slowdown in the headline figure, but is not enough to support a judgment that "U.S. consumers are broadly retrenching." Retail data are sales amounts not adjusted for inflation, and their resilience ultimately still needs to be converted into real consumption growth through corresponding price indicators. Components showed clear divergence: among 13 industries, 7 grew, 5 declined and 1 was flat. Other retailers rose 4.7%, health and personal care and gasoline stations each rose 1.5%, non-store retailers rose 1.2%, building materials rose 0.5%, food services rose 0.2%, and general merchandise stores rose 0.1%; furniture fell 2.3%, autos and parts fell 1.6%, clothing and sporting goods, hobby, book and music stores each fell 0.7%, food and beverage stores fell 0.2%, and electronics and appliances were flat. Large-ticket, interest-rate-sensitive consumption was weaker, while online channels and some everyday spending were more solid. Household purchasing power is also stratified. Brian LeBlanc, head of economic analysis at PNC Financial, noted that active credit card account balances grew more than 9% year-over-year, and younger households have thinner savings buffers and greater reliance on credit. This means retail resilience can support income expectations, but whether retailers can protect profits still depends on promotional intensity, customer mix and the ability to pass on costs. U.S. macroeconomic growth gets a floor, while equity market valuations remain constrained by interest rates Immediate trading after the data release showed that rate pressure temporarily outweighed growth positives. As of 8:44 a.m. Eastern Time on November 17, around 22:00 Beijing time, the two-year U.S. Treasury yield rose to 4.849% from about 4.82% before the data release, and the ten-year yield rose to 5.14% from 5.12%; the pre-market decline in the SPDR S&P 500 ETF Trust (SPY.US) widened from 0.1% to 0.2%, while the gain in the Invesco QQQ Trust (QQQ.US), which tracks the Nasdaq 100 Index, narrowed from 0.3% to 0.2%. Nasdaq-related assets were still attempting to repair, but the marginal reaction after the retail release was higher yields and stocks giving back some performance. This reaction has a clear macroeconomic basis: October CPI and core CPI rose 0.4% and 0.3% month-over-month, respectively, and consumer resilience leaves the Fed lacking a reason to pivot to easing due to rapidly weakening demand. Before the retail release, the market had priced in about a 95.4% probability of a 25 basis point rate hike in December; the dollar index had already risen above 103.50 during the Asian session. For the equity market, income expectations are supported, but the discount rate for forward profits and financing costs remain high, especially constraining AI infrastructure companies whose cash flows materialize later and which depend on continued financing. The adjustment magnitude facing this repair round remains large. The Philadelphia Semiconductor Index rebounded about 17.2% from November 5 to 11, then fell about 9.0% through the 16th; on Monday it fell 5.28% in a single day, while the Nasdaq 100 fell 2.67%. The Philadelphia Semiconductor Index has cumulatively retraced about 34% from its June high and remains deep in a technical bear market. Asia has also not escaped pressure: Taiwan's weighted index and South Korea's KOSPI plunged about 6% and 3.6% on Monday, respectively, and fell another 2.31% and 2.59% on Tuesday. Leadership transitions, export restrictions and doubts about AI investment returns all played a role; this selloff cannot simply be attributed to the end of the AI infrastructure boom or pessimistic expectations of recession. Consumption has not stalled, and the AI theme remains on a re-rating trajectory: funds are beginning to distinguish between economic resilience and actual AI computing returns The Bank of America survey precisely reveals this distinction: fund managers reduced tech allocations but did not broadly bet on recession. The survey conducted November 612 showed that the net overweight ratio for tech stocks fell from 37% to 12%, a decline of 25 percentage points and the largest single-month drop since 2006; this measures the difference between the proportion of respondents overweight and underweight. Consumer resilience can support corporate revenue, but has not simultaneously eased the interest-rate constraint on highly valued stocks: after the data release, two-year and ten-year Treasury yields moved higher, and the Nasdaq 100 ETF gave back part of its pre-market gains, reflecting that the market is still weighing growth support against financing costs. At the same time, the Bank of America survey showed that the net overweight ratio for tech stocks fell from 37% to 12%, the largest single-month decline since 2006, but investors have still not shown sufficient panic. The resilience of U.S. consumption and the market re-pricing of popular AI computing-related stocks are occurring simultaneously: the former reduces fears of a rapid stall in aggregate demand, while the latter still depends on whether computing power procurement, project delivery and customer collections can support forward profits. Whether the U.S. equity repair can continue hinges on whether improved corporate earnings can offset interest-rate pressure, and whether AI growth expectations can be validated by sustained cash flow. The Bank of America survey showed that only 3% expect a recession in the next year, and the cash ratio fell to 3.6%; a net 38% believe companies are overinvesting, and 52% expect AI hyperscalers to announce capital expenditure cuts in the next six months. The Bank of America strategist team led by veteran Wall Street strategist Hartnett therefore still believes that positioning cooling has not yet reached a level of panic sufficient to support contrarian buying. Strong consumption and reduced tech allocations can coexist, because funds are re-rating the returns on AI investment rather than comprehensively rejecting the U.S. economy. The target price adjustments for NVIDIA Corporation further reflect this re-pricing. Goldman Sachs Group, Inc., Citi and UBS Group AG respectively cut their target prices from $300 to $220, from $285 to $225, and from $300 to $240, all maintaining Buy ratings. Goldman Sachs Group, Inc. emphasized the impact of leadership uncertainty on valuation multiples; UBS Group AG also lowered its 2028 revenue forecast, reflecting potential constraints on business scale from factors such as export policy. Being bullish long term and lowering the price one is willing to pay are not contradictory, and consumption data cannot directly eliminate these company-level discounts. Well-known semiconductor supply chain research firm SemiAnalysis cut its 2028 AI accelerator capital expenditure forecast by 28% to $758 billion, cut its 2030 forecast by 40% to $923 billion, and lowered its 20262030 cumulative logic silicon wafer shipment area forecast by 17%. This is a revision to the previous growth path; the new model still expects 2030 spending to be significantly higher than in 2026. The latest AI supply chain investment logic shown in SemiAnalysis's latest forecast report is thatrevenue curves on the AI application side must cover computing costs, financing must support data center construction, and power access and system delivery determine when chip orders convert into revenue. SemiAnalysis emphasized that the latest forecast model adjustments involve revenue growth, financing and deployment constraints, and were not triggered solely by Jensen Huang's departure. SemiAnalysis cut its forecast for global AI accelerator spending in 2028 by 28% to $758 billion and in 2030 by 40% to $923 billion; but the new model still expects spending to continue growing from $392 billion in 2026. It should be noted that what has been cut here is the previously expected expansion path, and the scope is concentrated on accelerators and cannot be expanded to all data center capital expenditure. The institution also expects that of the U.S. data center capacity originally scheduled to come online between the second half of 2027 and the first half of 2028, 5.8 gigawatts will be delayed, but emphasized that revenue growth and the financing environment are the more important reasons for the adjustment. Analysts at the institution said that, broadly consistent with the forecast model, investors are demanding that future procurement scale be re-matched with paid demand, bankability and actual delivery progress. Retail data improve the macroeconomic foundation for corporate revenue, while a sustained rebound in AI stocks still requires validation from profits and financing conditions. What deserves more attention now are retail and payment companies that can convert consumer resilience into cash flow, as well as AI suppliers with clear delivery arrangements, collection capacity and funding guarantees. If inflation gradually moderates and financing costs decline, consumer resilience will more easily joint improvement in earnings and valuations; the U.S. equity market performance after the latest month's retail and historical revision data, especially tech stock dynamics, still belongs to a phase in which tech stocks are attempting to repair while the valuation-level interest-rate constraint has not yet eased.