Proprietary measures of risk and investor anxiety persuade BofA that AI euphoria is no bubble yet

AI is unlikely to be overwhelmed by the rise in interest rates, strategists argue.

As global bond yields rise alarmingly, investors are becoming understandably concerned about whether it will derail seemingly unstoppable artificial-intelligence trade.

Bank of America is reassuring clients that bond yields have to spike much higher than presently to properly derail the AI trade. Furthermore, because earnings growth in the sector is so strong, it's outpacing share-price appreciation, meaning that stocks are actually seeing smaller price-to-earnings multiples. From the conclusions drawn by Bank of America's proprietary "bubble risk indicator," analysts remain relatively untroubled about U.S. stock indices and say equities will quickly bounce back from any pullback.

The comfort was provided in a report published Wednesday by Bank of America's global equity derivatives research team, led by Benjamin Bowler. Their bubble risk indicator is a price-based measure, designed to detect bubble-like asset dynamics with the model incorporating an asset's returns, volatility, momentum and fragility into a single reading.

History is what provides Bank of America's analysts with their confidence. In the late 1990s, they point out that U.S. long bond yields BX:TMUBMUSD30Y jumped more than 200 basis points and the Fed hiked more than 100 basis points as Nasdaq soared. During the dot-com bubble, stock prices decoupled very clearly from fundamentals and the chart below illustrates that this phenomenon is not present yet in 2026.

During the dot-com bubble, stock prices clearly decoupled from fundamentals, a phenomenon not yet seen.

Nor have bond market moves this year been anywhere near as scary as they have been in the past. The 30-year's maximum drawdown so far in 2026 was 7% compared to the sell-off of 25% in 2022. "Bond yields at multi-year highs may be sending a concerning fiscal signal" but the comparatively muted reaction of both bonds and equities suggests the associated tightening in financial conditions has been materially less disruptive so far.

The note also makes the point that U.S. interest payments as a percentage of GDP are roughly 4.1% whereas in the late 90s they had ballooned to a figure approaching 5%. AI adoption, and the efficiencies and productivity gains it promises, may provide the U.S. with a plausible path to absorbing higher interest-rate costs.

Significantly, despite the plethora of headlines about bond yields, volatility markets are not yet concerned, Bowler writes. He cites another of Bank of America's proprietary models, its global financial stress index that captures between 20 and 40 measures of market anxiety across five asset classes, and reveals that thus far, its latest reading shows no reaction. At present, derivative markets are sanguine, Bowler says.

Despite global long-dated yields hitting decade+ highs, Bank of America says its interest-rate stress measure has not reacted.

This may not always be so. The report warns that upcoming Fed meetings could catalyze rate volatility if the tightening that the markets demand isn't forthcoming. "The risk," Bowler stresses, "is that a seemingly benign policy decision produces a disorderly bond market tantrum by undermining confidence in the broader policy mix."

The note finishes by reiterating that bubble-like environments can chew through macro risk, "at least for some amount of time." The technological leap AI promises "can result in investors pulling forward the value of future growth and overcoming macro headwinds."

In Thursday trading U.S. 10-year bonds were yielding 4.83% and Nasdaq futures were indicating a lower open.

-Jules Rimmer