Dot-com and AI bubbles: Can data from the past be helpful to match the price bubble euphoria phase using dynamic time warping?

Abstract

The article investigates the existence of a price bubble in the artificial intelligence market, employing the Generalised Supremum Augmented Dickey-Fuller test and dynamic time warping methodology. It proposes a method to detect the end of the price bubble euphoria phase, generating an average profit of close to 7 % over 5 days and over 10.5 % over 20 days, with almost 90 % effectiveness. The study found that the AI market experienced a phase of euphoria due to the COVID-19 pandemic. Four companies – META PLATFORMS, WORKDAY, SALESFORCE and NVIDIA CORPORATION – were identified, whose stock prices currently resemble the euphoria phase.