Snapshot
Bank of America’s August Global Fund Manager Survey injects an unusual mixture of alarm and reassurance into market sentiment. On one hand, 32% of surveyed managers now view an AI bubble as the principal tail risk globally. On the other hand, some of the most crowded trades have already been reduced, and survey responses signal continued capital spending from major cloud providers. Taken together, the report portrays markets that are heavily positioned yet increasingly selective about where they will redeploy capital.
Headline survey results
- AI bubble ranks as the top tail risk at 32%, surpassing inflation and geopolitical concerns.
- Cash on the sidelines sits at 3.5% of portfolios - the sixth-lowest level recorded since 1998 and below Bank of America’s contrarian "sell" threshold of 4%.
- 56% of managers expect a "no landing" scenario for the economy, the greatest share since February 2022.
- Global equity allocation is net 56% overweight, the highest reading since November 2021.
Those figures point to a market with strong exposure and limited defensive cash cushions at a time when a single thematic concern - AI froth - sits at the top of risk lists.
Contradictory signals in positioning and spending
Not all survey indicators are uniformly gloomy. Semiconductor exposure, previously the most crowded trade, shows notable de-risking: long positioning in semiconductors fell from 82% to 53% in 30 days. At the same time, 71% of fund managers do not expect hyperscaler capital expenditure cuts this year, and Bank of America raised its forecast for hyperscaler AI capex to $3.6 trillion through 2028 - a 17% increase month-over-month. Those dynamics indicate durable demand for core AI infrastructure even as peripheral, speculative positions are being trimmed.
Regional nuance: Asia fund managers tilt defensive
The Asia-focused fund manager responses amplify the rotation narrative. In that cohort, 59% of managers report rotating into defensive, value and cyclical stocks to hedge against AI downside risk - a level more than double that recorded in July. Nearly two-thirds of Asian managers said they need clearer proof of AI monetization before increasing exposure to the theme. The regional results underscore a growing preference for portfolio hedging amid uncertainty over revenue conversion from AI investments.
Technical backdrop: Nasdaq 100 overbought but trending
The Nasdaq 100 (NDX) was at 29,995, roughly 2.5% below its all-time high, and exhibited Strong Buy signals across daily, weekly and monthly horizons. Several momentum indicators registered elevated, near-extreme readings:
- Daily StochRSI: Nasdaq 100 95.96, NVIDIA 94.09 - characterized as extreme overbought.
- Weekly StochRSI: Nasdaq 100 92.41, NVIDIA 93.52 - characterized as extreme overbought.
- Monthly RSI: Nasdaq 100 70.87, NVIDIA 71.3 - entering overbought territory.
- Monthly ADX: Nasdaq 100 43.34, NVIDIA 41.0 - indicating a strong trend.
Those readings frame a clear tension: trend strength is confirmed by ADX above 40 on the monthly chart, yet momentum oscillators imply the market is stretched. The survey notes that, historically, a monthly RSI above 70 on the Nasdaq 100 has often preceded 5-10% pullbacks within four to eight weeks, even during bull markets.
NVIDIA as the bellwether
NVIDIA Corporation (NVDA) was trading down 2.1% in premarket action at $220.38, a modest correction against the wider survey noise. The company’s near-term metrics cited in the survey remain robust:
- Revenue growth: 70.7% and still accelerating.
- Gross margin: 71.1%, signaling pricing power.
- Return on equity: 114.3%.
- Forward P/E: 24.9x.
- Fair value upside: 19.1%.
- Analyst target upside: 33.3%.
- Beta: 2.22.
Bank of America reiterated a Buy rating and set a $350 price target after meetings with NVIDIA management regarding a $105 billion OpenAI commitment. BofA’s sum-of-parts analysis implies NVIDIA trades at a roughly 50% discount to fair value on a calendar-year 2027 basis.
A framework for rotation
The survey’s conclusions do not amount to an outright call to abandon AI-related investments. Rather, it suggests the easy gains may be behind the most speculative names and that positioning has become crowded. The report outlines a set of considerations for investors contemplating rebalancing:
When not to rotate
- If you hold AI infrastructure leaders (for example NVIDIA, TSMC, hyperscalers) that trade at forward P/E below 30x and with revenue growth above 40%.
- If technology exposure is under 30% of your equity portfolio.
- If you have a time horizon of three years or more, given the scale of the $3.6 trillion capex cycle.
- If you can tolerate a 15-20% drawdown without being forced to sell.
When to consider partial rotation
- If you own speculative AI names without a clear monetization pathway - a concern echoed by nearly two-thirds of managers in Asia.
- If technology represents more than 40% of your portfolio - consider trimming toward 25-30%.
- If you are sitting on gains of 50% or greater in AI-focused holdings and have not realized profits.
- If headline risk is causing material stress or sleeplessness.
For those trimming exposure, the survey points toward rotation targets including value, cyclicals and defensive ETFs. Examples cited include iShares Core High Dividend (HDV), which was down 0.44% and yields around 3.5%; Invesco S&P 500 Low Volatility (SPLV), down 0.81%; and iShares Core MSCI International Developed (IDEV), down 0.17% for geographic diversification away from concentrated U.S. tech exposure.
The contrarian warning and a middle path
There is a paradox for contrarian investors. The most crowded trade has already been materially reduced from 82% to 53% in one month, meaning the pool of marginal sellers may be smaller than feared. Yet the low cash level at 3.5% triggers Bank of America’s Cash Rule sell signal - a red flag that occurs when investors are broadly fully invested and bullish.
The survey also records record-high expectations for a "no landing" outcome, which raises the risk that disappointed growth or profit expectations could leave heavily invested managers exposed. A middle-path recommendation in the report is tactical: lock in profits by trimming 15-20% of AI or broader tech exposure into strength, redeploy some proceeds into the defensive and value exposures highlighted by the survey, and retain core positions in infrastructure winners given the ongoing capex cycle. This approach aims to reduce speculative tail risk while preserving participation in a continuing rally.
Note: Numbers and survey readings above reflect the Bank of America August Global Fund Manager Survey and related technical and company metrics as cited in the survey output.