BofA Survey: 56% of Fund Managers View AI Stocks as in a Boom, Not a Bubble
Claire Weston
Bank of America's June fund-manager survey shows 56% of respondents call AI stocks a boom, yet 80% flag long global semis as the most crowded trade ever recorded — institutions are staying long while systematically trimming exposure, a contradiction that marks a pivotal window for the AI rally.
Boom or bubble — how did fund managers vote?
56% say AI stocks are in a "boom" phase — momentum still building, FOMO pulling fresh capital in — while only 21% label it "euphoria."
This means → most allocators believe the rally has not peaked, but acknowledge the late-cycle risk is closing in.
A further 9% say the profit-taking phase has begun; 12% separately flag long Magnificent 7 (Apple, Microsoft, Nvidia and four other mega-cap tech names) as a crowded trade.
"Most crowded trade" hits an all-time high — how crowded?
80% of respondents call "buy and hold global semiconductors" the most crowded trade right now — the highest reading in the survey's history.
In plain terms = eight out of ten managers know everyone is in the same boat, yet most have not stepped off.
This reflects an unprecedented level of awareness around concentration risk — vast sums piled into a handful of names.
Where is the money moving — what do positioning shifts reveal?
Global equity net overweight fell from 50% to 38% month-on-month (down 12 pp); tech net overweight dropped from 33% to 26%; average cash rose from 3.9% to 4.1%.
Strategist Michael Hartnett described the move as "taking some chips off the table before summer."
This means → institutions are not turning bearish — they are tactically de-risking at highs. Trimming, not exiting.
Only 9% of investors believe AI's positive impact on equities is fully priced in, suggesting most still see the long-term story as intact.
What scares them most — how has tail-risk perception shifted?
34% rank "a second wave of inflation" as the top tail risk, holding first place.
"AI bubble" jumped from 5% to 28% in just two months, vaulting to second place.
In plain terms = almost nobody worried about an AI bubble two months ago; now nearly three in ten managers treat it as a top threat — the risk narrative is repricing fast.
Geopolitical-conflict concern collapsed from 44% to 12% over the same period. The market's fear center has migrated from external shocks to internal valuations.
How is Asia-Pacific splitting — who is hedging, who is riding it out?
On AI-trade downside risk, Asia-Pacific managers split in a rare dead heat: 41% choose not to hedge and stay overweight; another 41% are de-risking — shorting crowded AI beneficiaries, rotating into other parts of the AI value chain, or shifting to defensives.
In June, Asia-Pacific (ex-Japan) flows moved out of materials and consumer discretionary into financials and telecoms. Tech hardware overweight now exceeds semis; financials have become the new favored sector.
41% of respondents still see Taiwan as the biggest AI-cycle beneficiary. In Japan, earnings expectations are the key equity driver; tech hardware and banks are tied as the second-most-favored sectors, behind semis.
What is the core tension — can "trim but stay" last?
The current picture: FOMO and crowding anxiety coexist. Institutions maintain overweights while systematically cutting exposure.
This means → the market has not rejected the long-term AI narrative, but patience for near-term earnings delivery is narrowing.
In plain terms = the whole game hinges on one variable: whether AI companies can deliver earnings fast enough to justify current valuations — fast enough and the boom extends; too slow and the crowded trade unwinds.
Content is for reference only, not financial advice.