US equities positioning, flows, and market observations Report Interpretation

Overview

This weekly desk publication combines US-equity market performance, client positioning, derivatives activity, sector observations, commodity positioning, and a launch of single-stock hedge baskets. Its central message is that AI-linked technology demand and flow data remain constructive, but the market is navigating a more hawkish rate backdrop, geopolitical uncertainty, and substantial dispersion across sectors and themes.

Core views

US stocks ended the week broadly unchanged, with the S&P 500 little changed, as investors balanced a hawkish Federal Reserve rate hike against stronger growth signals and no clear evidence that the AI-infrastructure investment cycle is fading. Bitcoin Sensitive, AI Software and Agentic AI were among the strongest thematic baskets, gaining 8.83%, 6.84% and 4.94% for the week, respectively. By contrast, IPPs, Global Rare Earths, Natural Gas, Power Infrastructure and Alternative Asset Managers declined. The desk also notes an early-week rotation away from AI-linked shares and broader momentum after comments concerning the pace of frontier AI development, followed by a recovery into week-end. Prime Services flow data showed the largest US-equity net buying in five weeks, at +0.9 standard deviations relative to the prior year. Macro products accounted for 60% of total dollar net buying, primarily through short covering, while single stocks represented 40% and were bought for a third consecutive week. In single stocks, long buying exceeded short sales by 1.5 to 1, and seven of 11 sectors were net bought. Information Technology, Communication Services, Financials and Utilities led purchases, whereas Health Care, Energy and Industrials were the most sold. Technology and Communication Services were the most net-bought sectors for a third consecutive week; software, semiconductors and semiconductor equipment, telecom services, IT services, and interactive media led the TMT buying. Software allocation rose to 5% of total US net market value, from a February record low of 1.3%, although it remained below the 7% level at the start of the year. Health Care was the principal countertrend: hedge funds sold the sector for a fifth straight week and at the fastest pace in nearly six months, at -1.8 standard deviations versus the prior year. Biotech led the selling, followed by health-care equipment and supplies and providers and services. The sector long/short ratio fell to 2.32 from a March year-to-date high of 3.04, its lowest since July 2025. Separately, Health Care outperformed during the week, with XLV up 1.8% versus SPY down 20 bp, as AI-related concerns contributed to a rotation toward more defensive end markets. Sector commentary nevertheless remained uneven, including ongoing caution on managed care, mixed biotechnology developments, and uncertainty around clinical and M&A catalysts. The macro backdrop became less supportive for rate-sensitive sectors. The Fed raised the policy rate by 25 bp to 3.75%-4.00%; its projections indicated one further hike in 2026 and no cut in 2027, alongside slightly higher inflation projections. Goldman Sachs Global Investment Research expected another hike at the October meeting, while maintaining a 3.25%-3.5% neutral-rate estimate and adding a third projected cut in March 2028 alongside September and December 2027 cuts. Financials and Real Estate underperformed the S&P 500 as investors considered whether higher rates could restrain growth and capital expenditure. The desk flags Net Lease funding costs and leverage, potential pauses in CRE-broker activity, and the interaction of rates with office and industrial leasing tailwinds. Financial-sector conference feedback was largely better than feared on deposit costs, slightly negative on lending spreads, benign on credit quality, and positive on alternative managers' participation in the AI-capex cycle. Within technology, the desk characterizes recent conference feedback as bullish for Tech and AI, but says investors remain focused on what can restore momentum in the AI-infrastructure trade. It highlights an emerging “Inference Economy” theme comprising asset-light, AI-inference-exposed cybersecurity and data-infrastructure companies. The reasoning is that these businesses may grow through potential capex fluctuations or be less exposed to changes in pricing and supply-demand conditions. The desk observes that investors currently appear more rewarded for broad directional exposure to data infrastructure or security than for identifying specific share gainers, although it acknowledges debates over entry points and valuations. Derivatives flows were tactical amid geopolitical risk, rate uncertainty and AI-safety concerns. Total US option volume was 7.14% above its 200-day moving average as clients used options for protection against both tails. Thursday's rally generated the largest one-day decline in Goldman Sachs's Vol Panic Index in ten years, and the third-largest decline in its data history back to 2006. Implied volatility continued to compress in TMT; Nvidia implied volatility reached multi-year lows, prompting the desk's preference for outright calls as a stock-replacement approach with defined downside. The GSTMTMOM momentum basket fell 11.53% on Monday before recovering 2.08%, 3.98%, 2.56% and 2.97% over the following four sessions; over three months it was more than eight times as volatile as the S&P 500. The report also introduces more than 1,000 flagship single-stock hedge baskets as alternatives to broad market or sector proxies. Goldman Sachs uses AI-powered thematic mapping, fundamental analysis, and MKTMediaStats data to identify relevant peers and end markets, then systematically optimizes baskets for thematic relevance, liquidity, diversification and borrow cost. Baskets are periodically rebalanced as competitive landscapes, business models and exposures change. The Applied Digital hedge basket, GSSHAPLD, covers bitcoin miners, AI-infrastructure providers, data-center operators, networking-equipment companies and semiconductor beneficiaries. It is described as having a 5% maximum weight and capacity to trade $100 million per day at 10% ADV; it recorded 93% correlation with APLD versus 74% for the Nasdaq 100. A separate construction description cites 22 peers, approximately 92% correlation versus 67% for NDX, about 6% maximum weights, and notes that the basket underperformed both APLD and NDX on a normalized basis. The Pan African Resources basket, GSSHPAFL, had 91% historical correlation to PAF LN versus 62% for SXPP. In futures and commodities, major US-equity index calendar spreads cheapened into expiration after an initially richer roll. December ES funding had settled around 50-60 bp earlier in the year, but the TRF curve still priced a year-end premium, with Z6/F7 funding at FF+165 versus about FF+190 before the roll. Brent positioning had become more bullish as geopolitical disruptions lifted prices and open interest, but the report notes subsequent liquidation when pipeline-restoration expectations reduced perceived outage risk. Gold fell roughly 10% from its August high to $4,273 per troy ounce after the Fed hike before finding support above $4,300; GCZ6 rebounded 3.4% from its lows, and the desk viewed the pullback as an opportunity to rebuild length, citing intact structural support and 32 bp of additional tightening already priced by year-end.

Analysis framework

The publication combines market-price and thematic-basket performance, Prime Services client flow and positioning data, sector-specialist observations, derivatives positioning, and futures and commodity market measures. For its hedge baskets, Goldman Sachs maps company themes and peer exposures, then optimizes constituent selection subject to relevance, liquidity, diversification, borrow-cost and concentration constraints before monitoring correlation and rebalancing periodically.

Methodology notes

• Quantitative, Factor, and Portfolio Theory

  Thematic similarity mapping and systematic basket optimization

  Goldman Sachs uses MKTMediaStats and fundamental analysis to identify peers and end markets linked to a company, then optimizes a liquid and diversified hedge basket intended to track those exposures.

• Quantitative, Factor, and Portfolio Theory

  Historical correlation and tracking comparison

  The report compares hedge-basket correlations with the relevant stock against conventional benchmarks to assess how closely the basket historically tracked company-specific exposure.

• Event-Driven and Behavioral FinanceFund-Flow and Positioning Analysis

  Prime Services flow and positioning analysis

  The desk interprets net buying and selling, long versus short activity, leverage, allocation and percentile measures to describe investor positioning and changes in sentiment.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Applied Digital (APLD) / GSSHAPLD

  Goldman Sachs presents GSSHAPLD as a targeted hedge basket designed around Applied Digital's compute, power and digital-infrastructure demand exposures.

  Strengths

  93% correlation to APLD versus 74% for the Nasdaq 100; diversification, liquidity constraints and exposure to 22 aligned peers are cited.

  Weaknesses

  The basket underperformed APLD and NDX on a normalized basis.

  Comparison

  A cited comparison shows approximately 92% correlation for GSSHAPLD versus 67% for NDX; another shows 93% versus 74% for the Nasdaq 100.

  Risks

  Basket composition, correlation and thematic exposure can change; the report notes the need to review composition before trading.

• Pan African Resources (PAF LN) / GSSHPAFL

  Goldman Sachs presents GSSHPAFL as a hedge basket linked to gold-mining, diversified-mining and precious-metals drivers associated with Pan African Resources.

  Strengths

  91% historical correlation to PAF LN with liquidity and diversification constraints.

  Comparison

  The reported 91% correlation compares with 62% for SXPP.

Key data

• US equity net buying+0.9 SDs versus one-year historyLargest net buying in five weeks; driven by short covering in macro products and long buying in single stocks.
• US fundamental long/short gross leverage209.4%Up 2.5 percentage points week on week; 33rd percentile over one year.
• US fundamental long/short net leverage49.9%Down 0.3 percentage points week on week; seventh percentile over one year.
• Health Care hedge-fund selling-1.8 SDs versus one-year historyFifth consecutive week of net selling and fastest pace in nearly six months.
• Federal Reserve policy rate3.75%-4.00%Raised 25 bp at the September FOMC meeting.
• GSSHAPLD correlation to APLD93%Compared with 74% for the Nasdaq 100 in one cited comparison.
• US option volume+7.14% versus the 200-day moving averageReflects tactical use of options amid geopolitical, rates and AI-safety uncertainty.

Impact & implications

The report portrays a market in which broad equity performance masks meaningful rotation: AI-linked technology and inference themes retain flow and valuation support, while Health Care positioning has weakened and higher rates create a more difficult backdrop for Financials and Real Estate. The hedge-basket launch is positioned as a way to obtain more company-specific peer and end-market exposure than conventional sector or index hedges.

Risks

• Geopolitical risk, rate uncertainty and emerging AI-safety fears contributed to tactical client positioning and elevated options activity.
• Higher rates may weigh on economic growth and the broader capital-expenditure cycle, while increasing funding-cost and leverage concerns for rate-sensitive real-estate businesses.
• The report notes uncertainty around AI-infrastructure trade momentum and debates over technology entry points and valuations.
• The GSSHAPLD basket underperformed APLD and NDX on a normalized basis despite its higher historical correlation.

What to watch

• The Trump/Xi summit, described as Xi's first US trip since 2023.
• The S&P 500 implied move of 1.29% through September 25.
• Investor days for NVO, ONON, FSLY, PATH, MCD, AUR and SHW.
• META's Connect Developer Conference, including Zuckerberg's Wednesday presentation.
• Whether AI-infrastructure momentum returns and whether capex-light inference themes continue to receive valuation support.
• Rate-market developments, their implications for growth and capex, and their effect on Financials and Real Estate.