Cornell Study Finds AI Market Is Not in a Bubble, But Select Tech Stocks Are Overvalued

While widespread financial market debate has centered on whether artificial intelligence is fueling an industry-wide asset bubble, new academic research from Cornell University indicates that the broader AI sector is not in a bubble. Instead, speculative pricing dynamics are isolated to specific public technology corporations, including Google parent Alphabet Inc., according to an empirical analysis of equity trading patterns.

The findings were detailed in a study titled "Is There an AI Bubble? Robust Data-stamping for Periods of Exuberance," published in the journal Frontiers in Mathematical Finance and first reported by TechXplore. Co-authored by Martin Wells, a professor of statistical sciences at Cornell University, and doctoral student Abir Sarkar, the research examined daily stock price movements from 2020 through April 2026 across major tech firms, AI hardware suppliers, semiconductor makers, and cryptocurrencies.

To evaluate asset pricing, the researchers developed a statistical model called the Stochastic Volatility-robust Augmented Dickey–Fuller (SV-ADF) framework. Conventional bubble-detection models often label entire technology sectors as systemic bubbles during periods of market turbulence. In contrast, the SV-ADF methodology filters out transient price volatility to isolate individual equities exhibiting true speculative exuberance.

"We hear the term 'bubble' used frequently, but the more important question is whether we can identify and quantify the underlying characteristics associated with one," Wells said. "Our analysis suggests that some segments of AI-related companies exhibit characteristics consistent with speculative bubble dynamics, while others do not. Treating 'AI' as a single category therefore obscures important heterogeneity across firms and sectors."

Sarkar emphasized that the granular model prevents broad-brush mischaracterizations of entire market segments. "Our method is not saying that everything is a bubble. That is the key distinction from the standard bubble-detection methods," Sarkar said. "Our method is robust enough to identify transient volatility spikes and is sufficiently selective to distinguish bubble episodes at the individual-stock level, even within closely related industry groups."

In its firm-specific analysis, the study identified Alphabet as demonstrating clear statistical evidence of overvaluation. Alphabet's share price surged over 70% in the preceding 12 months, significantly outperforming the Nasdaq Composite's roughly 20% growth over the same span. Looking historically, the researchers found that virtually all semiconductor companies entered bubble territory following the November 2022 release of OpenAI's ChatGPT, while electric vehicle maker Tesla showed bubble behavior in 2020. In digital assets, Bitcoin and Ethereum exhibited exuberance starting in December 2020. Though those previous bubbles eventually collapsed, isolated exuberance later re-emerged in specific cases.

The study evaluated a broad set of tech assets, including the "Magnificent Seven"—such as Apple, Amazon, Nvidia, and Meta—alongside semiconductor and infrastructure leaders TSMC and Broadcom. Wells and Sarkar plan to extend their research alongside Robert Jarrow, an investment management professor at the Cornell SC Johnson College of Business. "Identifying stocks with bubbles is important because bubbles eventually burst, which results in dramatic price declines," Jarrow said. "Knowing which stocks contain bubbles enables investors to make better investment decisions."