Fund-manager optimism remains strong, but September positioning shows a more cautious turn amid bond-yield and AI-capex risks.

Overview

This monthly global fund-manager survey tracks investor macro expectations, perceived risks, and cross-asset, regional and sector positioning. September respondents remained broadly constructive on growth and AI spending, but their positioning became less aggressive as concern shifted toward bond yields, corporate overinvestment and potential credit stress tied to AI hyperscaler capital expenditure.

Core views

The September survey indicates that optimism has moderated rather than disappeared. BofA's broad sentiment measure—combining cash, equity allocation and global growth expectations—fell to 7.0 from 8.0, making September the least bullish reading in three months after August was the third-most bullish survey since 2022. Cash rose to 3.9% of assets under management from 3.5%, the largest monthly increase since March 2026. The firm's contrarian Cash Rule nonetheless remained in sell-signal territory because it signals “sell” at cash of 4.0% or below; the report says a return to the 4–5% neutral cash range would make it safer to increase risk exposure. Macro expectations remain resilient. Net 8% expected stronger global growth over the next 12 months, down from 14% in August. A 55% majority expected a “no landing,” slightly below August's 56% record high; 38% expected a soft landing, up from 34%, and 2% expected a hard landing. However, the survey's alternative growth/inflation framing shows a more divided backdrop: 50% described the likely environment as stagflation and 38% as boom, while goldilocks and stagnation each drew 4%. Inflation expectations reversed, with net 4% expecting lower global CPI in 12 months versus net 3% expecting higher CPI in August, yet net 36% expected higher short-term rates, the highest reading since September 2022. Net 25% judged global monetary policy too stimulative, also the highest since September 2022. Policy and rates are central to the survey's risk assessment. Investors saw the Federal Reserve as behind the curve, and for the first time since September 2022 expected a flatter yield curve. On whether the Fed would raise rates before the November US midterms, 52% said no, down sharply from 72% in August, while 41% said yes, up from 22%. Regarding the US Treasury's larger bond-buyback program from September 9 to November 4, 46% expected no effect on Treasury yields and 29% expected yields to rise. The leading tail risk was a disorderly rise in bond yields at 33%, up from 27% in August and replacing an AI bubble, which fell to 28% from 32%. The survey's Financial Market Stability Risks Indicator rose to -1.1 from -2.1 as business-cycle, monetary, protectionist and geopolitical risks increased month on month. AI remains both a foundation of optimism and a major vulnerability. The report says investor conviction in rapid AI capital expenditure remains strong: 79% did not expect an AI hyperscaler to announce a capex cut in 2026, up from 71% in August, while 14% did expect a cut. At the same time, a record net 33% said companies were overinvesting, matching the February 2026 high. AI hyperscaler capex was identified by 42% as the most likely source of a systemic credit event, up from 38% in August; government debt was second at 25%. Long global semiconductors remained the most crowded trade at 53%, followed by short Treasuries at 18% and long Magnificent 7 at 7%, highlighting concentrated positioning around the AI theme. Cross-asset positioning became somewhat less risk-seeking. Equity allocation declined to net 49% overweight from net 56% overweight, though it remained 0.9 standard deviations above its long-term average. Bond allocation fell to net 48% underweight from net 39% underweight, the lowest level since May 2022 and a 17th consecutive month of underweight positioning. Cash moved to equal-weight from net 2% underweight, while reported risk-taking shifted to net 2% below normal from net 4% above normal in August. Real-estate allocation fell to net 21% underweight from net 7% underweight and reached its lowest level since September 2025. Commodities remained net 19% overweight, down from 24% overweight. The survey says 27% would need US 30-year Treasury yields around 6% to shift overweight government bonds over the next 12 months, whereas 19% would need a major top in equity markets. Regional and sector allocations show a clear rotation. Investors were net 25% overweight US equities, down from 27%, but still 1.4 standard deviations above their historical average. Emerging-market equities rose to net 38% overweight from 34%, while eurozone equities flipped to net 5% underweight from net 6% overweight, Japanese equities moved to net 1% underweight from net 1% overweight, and UK equities stayed heavily underweight at net 35%. The report's contrarian trades are long UK stocks/short US stocks, long staples/short banks, and long small cap/short large cap. Sector flows favored financials, healthcare and industrials. Banks rose to net 34% overweight from 30%, insurance to net 3% overweight from net 10% underweight, and industrials to net 20% overweight from 15%. Healthcare rose to net 36% overweight from 29% and remained consistently overweight since January 2018. Technology was unchanged at net 30% overweight. Conversely, consumer discretionary moved to net 21% underweight from 12% underweight, staples to net 33% underweight from 19% underweight—the largest underweight since January 2004—and investors also reduced REIT and telecom exposure. Relative to history, banks, US equities and healthcare were the most overweight positions, while staples, UK equities and REITs were the most underweight.

Analysis framework

The report aggregates September 4–10 survey responses from 190 panelists managing $512 billion of assets, including 170 respondents managing $470 billion who answered global questions. It compares net investor responses and allocations with prior months, historical extremes and long-term standard-deviation readings, then uses the results to identify macro expectations, crowding, tail risks and possible contrarian positioning opportunities.

Methodology notes

• Quantitative, Factor, and Portfolio Theory

  Fund-manager survey positioning and sentiment indicators

  BofA combines reported cash levels, equity allocation and growth expectations into a sentiment measure, and uses survey net balances to track expectations and positioning.

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

  Crowded-trade and contrarian analysis

  The report identifies the trades with the highest reported investor participation and presents contrarian long/short pairings against extreme positioning.

• Quantitative, Factor, and Portfolio Theory

  Historical z-score comparison

  Current allocations are compared with their long-term averages in standard deviations to show how unusual positioning is.

Asset mapping & comparison

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

• Global equities

  Investors remained overweight, though reduced the allocation in September.

  Strengths

  Net 49% overweight despite a monthly decline.

  Weaknesses

  Sentiment and risk-taking both moderated.

  Comparison

  Allocation fell from net 56% overweight in August.

  Risks

  Disorderly higher bond yields and potential AI-capex credit stress.

• Government bonds

  Investors maintained a pronounced underweight despite higher-yield risks.

  Strengths

  Some respondents would shift overweight if US 30-year Treasury yields reached around 6%.

  Weaknesses

  Net 48% underweight, the lowest allocation since May 2022.

  Comparison

  Underweight widened from net 39% in August.

  Risks

  A disorderly rise in bond yields was the leading reported tail risk.

• Global semiconductors

  Most crowded long trade, linked to sustained AI-capex conviction.

  Strengths

  79% did not expect an AI hyperscaler capex cut in 2026.

  Weaknesses

  Positioning is highly crowded at 53%.

  Comparison

  Crowding was unchanged from August.

  Risks

  AI hyperscaler capex was the most-cited source of a systemic credit event.

• Banks and insurance

  September rotation favored financials.

  Strengths

  Banks were net 34% overweight and insurance net 3% overweight, both their highest overweight since November 2025.

  Comparison

  Banks rose from net 30% overweight; insurance flipped from net 10% underweight.

  Risks

  The report's contrarian trade includes long staples/short banks.

• Healthcare

  A favored sector in September's rotation.

  Strengths

  Net 36% overweight; positioning has been consistently overweight since January 2018.

  Comparison

  Up from net 29% overweight in August.

• REITs and real estate

  Investors reduced exposure to real estate and REITs.

  Weaknesses

  Real-estate allocation was net 21% underweight, the lowest since September 2025.

  Comparison

  Underweight widened from net 7% in August.

  Risks

  Higher bond yields are the survey's leading tail risk.

Key data

• Survey panel190 panelists; $512bn AUMSurvey conducted September 4–10, 2026; 170 respondents with $470bn AUM answered global questions.
• Broad FMS sentiment7.0Down from 8.0 in August.
• Cash level3.9% of AUMUp from 3.5%; largest monthly increase since March 2026.
• Global economic outcome55% no landing; 38% soft landing; 2% hard landingViews on the next 12 months.
• Biggest tail risk33% disorderly rise in bond yieldsUp from 27% in August; AI bubble was 28%.
• AI hyperscaler capex credit risk42%Respondents identifying it as the most likely source of a systemic credit event, versus 38% in August.
• Equity allocationNet 49% overweightDown from net 56% overweight; 0.9 standard deviations above the long-term average.
• Bond allocationNet 48% underweightDown from net 39% underweight; lowest since May 2022.
• Most crowded trade53% long global semiconductorsUnchanged from August.
• Healthcare allocationNet 36% overweightUp from net 29% overweight.

Impact & implications

The report portrays a market still supported by expectations of resilient growth and ongoing AI investment, but with diminishing excess bullishness. Its survey evidence points to greater sensitivity to higher yields, concentrated semiconductor and AI positioning, and capital-spending risks, alongside a rotation toward healthcare, banks and industrials and away from rate-sensitive real estate and defensive staples.

Risks

• A disorderly rise in bond yields was the largest reported tail risk, cited by 33% of respondents.
• AI hyperscaler capital expenditure was cited by 42% as the most likely source of a systemic credit event.
• A record net 33% of respondents said companies are overinvesting.
• Long global semiconductors was the most crowded trade at 53%, indicating concentrated positioning.
• Respondents identified rising business-cycle, monetary, protectionist and geopolitical risks in the Financial Market Stability Risks Indicator.

What to watch

• Whether cash rises back into the 4–5% neutral range identified by the report as a safer point to increase risk exposure.
• Changes in short-term-rate expectations, yield-curve views and the response of Treasury yields to the buyback program.
• Whether an AI hyperscaler announces a 2026 capital-expenditure reduction.
• The progression of bond-yield stress and its effect on asset allocation.
• Changes in US midterm expectations and the market reactions respondents associate with a Democratic sweep.
• Whether sector rotation toward financials, healthcare and industrials continues, or contrarian positioning in UK equities, staples and smaller caps gains support.