Does Yardeni Research believe there is an AI stock market bubble?
Yardeni Research believes that artificial intelligence-related stocks are showing characteristics of a market bubble, although its latest analysis suggests that the conditions associated with a major market correction are not yet fully established.
In a research note published this week, the firm introduced an AI Bubble Meter, a new indicator designed to assess both the existence of speculative market conditions and the likelihood that those conditions could lead to a downturn.
The indicator was developed using Anthropic’s Claude Fable 5.1 artificial intelligence model and incorporates historical comparisons with previous periods of elevated market valuations.
“The meter says there’s a bubble, but it’s not about to burst,” Yardeni Research stated.
How does Yardeni’s AI Bubble Meter work?
The AI Bubble Meter consists of two separate indicators, referred to as the Conditions dial and the Triggers dial.
The Conditions dial measures whether the characteristics of a financial bubble are present. Its framework draws on conditions observed before the technology stock market decline of 1999–2000 and the US housing market downturn of 2006–2007.
The Triggers dial evaluates whether the economic and financial developments that contributed to the end of previous market expansions are emerging.
These include rapid credit growth and sustained increases in the cost of capital, both of which can affect asset valuations and financing conditions.
By separating underlying market conditions from potential catalysts for a correction, the indicator is intended to distinguish between the existence of a bubble and the immediate risks associated with it.
What are the current readings of the AI Bubble Meter?
Yardeni Research reported a Conditions reading of 83 out of 100, indicating that several characteristics associated with previous financial bubbles are present.
The Triggers dial, however, stands at 34, suggesting that the factors historically associated with the end of speculative market expansions are less developed.
The firm identified an increase in short-term US Treasury yields as one emerging trigger.
The yield on the two-year US Treasury note has risen by 115 basis points from its December low, reflecting an increase in borrowing costs.
Nevertheless, Yardeni Research said that credit conditions outside AI-related sectors do not currently show the broader signs of excessive expansion observed before some previous market downturns.
“The preconditions of a bubble are in place and have been for a while, but the forces that ended past booms are only partially present,” the firm said.
What does historical market data suggest about the risk of a downturn?
Yardeni Research examined historical market conditions dating back to 1965 to assess how different combinations of its two indicators have been associated with subsequent stock market performance.
According to the firm’s analysis, periods with elevated Conditions readings but relatively low Triggers readings were followed by an S&P 500 bear market within two years in only 9% of cases.
During those periods, the index recorded a median inflation-adjusted gain of 18.3% over the subsequent two years.
The results differed when both indicators were elevated.
Yardeni Research found that periods characterised by high readings on both dials were followed by a bear market within two years in 42% of cases, with a median inflation-adjusted loss of 12.5%.
These figures represent historical associations identified by the firm’s model rather than forecasts of future market performance.
Why are US Treasury yields important to the AI bubble assessment?
Changes in Treasury yields are relevant to the firm’s analysis because rising borrowing costs can affect corporate investment, financing availability and equity valuations.
Yardeni Research is monitoring whether the two-year Treasury yield continues increasing towards the cumulative rises of approximately 200 to 350 basis points observed before the market reversals associated with the 2000 technology bubble and the 2006 housing bubble.
The current increase of 115 basis points remains below that historical range.
A sustained rise in yields could contribute to higher readings on the Triggers dial, particularly if accompanied by a broader expansion in credit risk.
“But the Triggers dial is the one to watch: If it starts rising toward High, so do the risks,” Yardeni Research said.
What will Yardeni Research monitor in upcoming technology earnings?
In addition to Treasury yields and credit conditions, Yardeni Research plans to examine third-quarter earnings reports from major hyperscale technology companies.
The firm will look for evidence of unused computing infrastructure or excess capacity, which could provide information about the relationship between AI investment and demand for computing services.
Hyperscalers have been investing in data centres and other infrastructure to support artificial intelligence workloads.
Signs that installed capacity is not being fully utilised could become relevant to assessments of future capital expenditure and the financial returns associated with AI infrastructure investment.
For now, Yardeni Research’s indicator suggests that the characteristics of an AI-related stock market bubble are present, while the financial triggers historically associated with major corrections remain only partially developed.