Wall Street Bets Big on AI at Six-Year Highs as Nearly Half of Fund Managers Predict a 2027 Crash

The most consequential opening morning of the Q3 earnings season is already underway. JPMorgan, Goldman Sachs, Wells Fargo, Citigroup, UnitedHealth and Johnson & Johnson are all reporting before the opening bell on Tuesday, October 13 — while Bank of America, Morgan Stanley and BlackRock follow on Wednesday, October 14. Every figure will be scrutinized against the same overriding question: is artificial intelligence generating real commercial returns, or is Wall Street running on spectacular borrowed conviction?

The October 2026 Paradox: Chasing Tech at Six-Year Highs

Bank of America's October 2026 Global Fund Manager Survey, released Tuesday morning, captures the central tension of this earnings season in a single data point: the proportion of fund managers running overweight positions in technology has risen to 37%, up from 30% last month, reaching its highest mark since August 2020. Investors are chasing the AI rally with more conviction than at any point since the post-pandemic rebound — yet roughly 45% of the same respondents expect a full AI bubble to materialize by 2027, with only 42% disagreeing.

BofA strategist Michael Hartnett has consistently flagged the contradiction between surging positioning and mounting valuation anxiety. The survey also comes one month after the September 2026 reading showed global equity overweight near 49%, suggesting that despite rising bubble fears, the rotation into risk assets — and into technology specifically — has only deepened heading into the most critical earnings stretch of the year.

The $570 Billion Question: Who Is Funding the Build-Out?

Alphabet, Amazon, Microsoft and Meta are collectively projected to spend $700 billion on AI infrastructure in 2026, and they are increasingly doing so through the bond market rather than from operating cash flows. Morgan Stanley forecasts AI-linked global debt issuance to approach $570 billion for the full year — nearly double last year's total pace, and running four times faster through May than the equivalent period in 2025. By 2027, hyperscaler capital expenditure is expected to surpass $1 trillion — a threshold that transforms AI infrastructure from a corporate decision into a capital markets dependency.

The scale of individual transactions underscores how structural these financing bets have become. Amazon raised C$14 billion in Canadian dollar bonds in June — the largest corporate offering ever in that currency — after closing a €14.5 billion euro-denominated deal in March, the largest euro corporate bond on record. Alphabet tapped yen markets as part of a multi-currency borrowing blitz that also included a 100-year sterling bond — the first century debt instrument from a technology company since Motorola in 1997. These are not routine financing decisions. They are structural wagers that AI revenue will materialize quickly enough to service obligations denominated in multiple currencies, measured across decades.

The Circular Trap Keeping Analysts Up at Night

The financing architecture of the AI boom contains a feedback loop that institutions are increasingly flagging. The largest AI labs require enormous compute capacity. The companies supplying that capacity — chip manufacturers, data center operators, cloud hyperscalers — are financed in part by the same institutional investors who are betting on those labs' success. A funding shortfall at a single major AI lab could reverse the chain in both directions at once.

Compounding the pressure: U.S. federal debt crossed $40 trillion for the first time in history on August 19, 2026, as rising interest payments and expanded defense spending outpaced revenue growth. Every basis point added to long-term Treasury yields tightens the screws on any leveraged borrower operating inside the AI financing stack — and the hyperscalers are now among the largest leveraged borrowers in the corporate bond market.

Dalio's Warning at the Right Moment

Speaking at the Forbes Global CEO Conference in Singapore on October 7, Bridgewater Associates founder Ray Dalio delivered what may be the most direct verdict yet from the world of institutional finance. He described the current AI market as "a classic bubble" approaching the moment when rising interest rates and the need to convert AI wealth into cash to service debt would force a sharp reversal. "I think we're close to that," Dalio said, describing conditions where the trigger hasn't fired yet — but where financing costs and leverage are visibly moving toward it.

Goldman Sachs strategists have offered the counterargument, maintaining that it remains premature to call a tech bubble given the robust earnings profile of AI's leading players — a disagreement between the bank's in-house view and the founder of the world's largest hedge fund that itself illustrates how unresolved this debate has become.

Earnings Season as the Stress Test

FactSet estimated S&P 500 earnings growth at 29.6% year-over-year for Q3 as of October 9 — the eighth consecutive quarter of double-digit gains — with Information Technology ranked among the top-performing sectors. Based on the historical pattern of companies beating estimates, the final realized growth rate is expected to exceed 35%. Goldman Sachs alone reported diluted EPS of $17.01 for Q3, crushing a consensus estimate of $13.65, as equities revenue surged 72%.

The decisive verdict, however, arrives the week of October 26–30, when Microsoft, Alphabet, Meta, Amazon and Apple all report — five companies representing roughly one-third of the S&P 500's total market capitalization. If those results cannot demonstrate that years of debt-financed infrastructure spending is converting into real commercial returns, the tension now visible in BofA's survey stops being a sentiment reading and becomes a pricing event. The Nasdaq 100 has gained approximately 22% year-to-date on the AI thesis. October will determine whether that gain reflects a genuine productivity revolution — or the opening act of an exceptionally expensive reckoning.

© {{Year}} Latin Times. All rights reserved. Do not reproduce without permission.