Valuing Digital Asset Treasury Companies: Integrating Augmented CAPM and Machine Learning in MicroStrategy

The integration of cryptocurrencies into corporate treasuries has created a new asset class of "Digital Asset Treasury" companies, its correlation with cryptocurrencies troubles traditional valuation frameworks. This study investigates the pricing mechanics of MicroStrategy as a case study, to determine whether its equity premium can be explained by standard market factors. Using daily data from 2020 to 2025, this study first identifies the return correlation between MSTR and BTC, which surged tor=0.695after BTC's appearance in its portfolio. This paper evaluates the limitations of the traditional Capital Asset Pricing Model (CAPM) and proposes an augmented CAPM incorporating a specific Bitcoin return factor. Linear regression results of the traditional model indicate that MSTR is a leveraged BTC proxy, while traditional CAPM only bringsR2=21.8%explanatory power. Furthermore, this study simulates the convexity of MSTR to BTC returns in extreme days using a non-linear function format. Finally, delivering enhancement for the model is necessary; thus, this study utilizes the Random Forest algorithm and SHAP analysis to rate all considered factors, and come up with a multi-factor, non-linear model specifically designed for the valuation of DAT companies.