The growing integration of graphical processing units (GPUs) into edge and cloud computingCloud computing infrastructure has made understanding their market dynamics increasingly important for dependable system design and resource planning. In recent years, global disruptions, including the COVID-19 pandemic and ongoing U.S.–China trade tensions, have contributed to significant volatility in the consumer GPU market, affecting the availability and pricing of products from leading vendors such as Nvidia and AMD. This volatility has introduced new challenges for infrastructure planners and system designers who rely on predictable access to computational resources. In this paper, we propose an intelligent modeling framework for forecasting GPU market share trends using machine learningMachine learning. Specifically, we develop two XGBoost regression models, augmented with principal component analysis (PCA)Principal Component Analysis (PCA) and lag features, to predict quarterly market shares for Nvidia and AMD based on GPU specifications and pricing data. The models achieve high predictive performance, with R2 scores of 0.977 for Nvidia and 0.965 for AMD, respectively, when tested on historical data from 1990 to 2024. By providing insights into future GPU availability and vendor competitiveness, our approach supports more informed and resilient planning in edge and cloud computingCloud computing environments.