BofA survey: 52% of fund managers expect an AI hyperscaler to cut capex within six months, up from 30%, as tech overweights fall the most since 2006.
- 52% of fund managers expect an AI hyperscaler to announce a capex cut within six months, up from 30% in October.
- Only 16% want companies to raise capital spending, the lowest share since May 2020.
- Net tech overweight fell to 12% from 37%, the largest one-month drop since June 2006.
Lead
A majority of global fund managers now expect one of the large AI infrastructure spenders to scale back investment. Bank of America's November Global Fund Manager Survey, conducted Nov. 6-12, found that 52% of respondents expect at least one AI hyperscaler to announce a capital expenditure cut within six months. The share was 30% in October. Only 16% want companies to increase capital spending, the lowest reading since May 2020. The results mark a sharp turn in sentiment toward ai stocks after a year in which data-center spending drove much of the earnings growth in US equities.
What Did the BofA Survey Show?
The survey shows that investors have moved from rewarding AI spending to questioning it. A record net 38% of respondents say companies are overinvesting, the highest reading in the survey's history on that question.
The panel included 193 fund managers overseeing $526 billion in assets, with 169 managers responding to global questions on $461 billion. Positioning moved quickly. The net share of investors overweight technology dropped to 12% from 37% in October, the biggest monthly cut since June 2006.
The survey's headline message was a call to pause AI capex, and the strategist who leads the survey described the mood as the end of boom and AI euphoria.
Why Are Investors Bracing for a Capex Cut?
Investors are bracing because the scale of spending now depends on debt markets and uncertain returns. Hyperscalers such as Microsoft (MSFT), Amazon (AMZN), Alphabet (GOOG) and Meta Platforms (META) have committed to multiyear data-center builds. Their size has pushed some of the financing onto bond markets and private credit.
That shift shows up in tail-risk rankings. An AI bubble is the top tail risk at 34% of respondents. An AI hyperscaler credit event follows at 31%. Among those who see a credit event as the main danger, 38% point to hyperscaler bonds and 30% to private credit.
The concern is less about demand for AI than about the return on capital. A cut by any one of the largest spenders would reduce orders across the supply chain, from accelerators to networking and power equipment. It would also set a precedent for peers that have so far matched each other's budgets.
How Are Investors Repositioning in AI Stocks?
Investors are trimming exposure but have not left the trade. Technology remains an overweight, and long semiconductors is still the most crowded position, though only 39% of respondents name it, down from 82% in July. That fall points to a gradual unwinding of the Nvidia (NVDA) and chip-supplier trade, not a capitulation. Long Magnificent Seven ranks lower at 12%, and short Chinese stocks is cited by 14%.
Cash levels average 3.6%, down from 3.9% in October, a low level by the survey's history that leaves limited buying power on the sidelines. Bond allocations improved, with the net underweight narrowing to 23%, a 13-point shift in one month.
Economic Backdrop
The economic outlook remains benign, which separates this episode from a broad risk-off move. About 50% of respondents expect a soft landing and 42% expect no landing. Only 3% forecast a hard landing. Investors are bullish on growth and corporate profits while doubting that current AI spending rates are sustainable.
That split explains the focus on capex. Investors expect the economy to hold up and are questioning whether companies are directing too much capital toward one theme.
What Comes Next for AI Spending?
The next test comes with quarterly earnings and 2027 budget guidance from the hyperscalers. Capex plans are typically updated on earnings calls, which makes the first-quarter reports a likely point for any revision. Companies that raise budgets will confirm the current trajectory. A cut or a slower growth rate would validate the 52% of managers expecting one, and it would have direct consequences for chipmakers and other suppliers.
Credit markets are the second channel. Wider spreads on hyperscaler debt or tighter private credit terms would raise financing costs for data-center projects and strengthen the case for restraint.
Outlook
The November survey records a reversal in how investors view AI infrastructure spending. Expectations of a hyperscaler capex cut have risen to 52%, support for higher spending is at a six-year low, and tech overweights have shrunk at the fastest pace since 2006. Positioning remains long technology and semiconductors, so the adjustment has further to run if spending guidance disappoints. Earnings calls and credit conditions will show whether the shift in sentiment becomes a shift in corporate budgets.
Mentioned tickers: MSFT, AMZN, GOOG, META, NVDA, BAC




