Blockchain-Driven Supply Chain Modernization to Strengthen U.S. Critical Infrastructure and National Economic Security

Abstract

The complexity and digitization of the supply chain industry in the US is making transparency, data accuracy, and operational efficiency more of a concern for all involved. To address this challenge, the integration of blockchain into machine learning (ML) has been proposed as a viable approach to support secured data sharing and intelligent decision making. In the backdrop of the above-mentioned issues, this paper investigates how blockchain and machine learning technologies have been adopted and discusses their effects on supply chain transparency and operational efficiency in US-based supply chains. A quantitative, machine learning approach is implemented with cross-sectional survey data from 300 US supply chain professionals in manufacturing, logistics, and retail. The dataset contains blockchain adoption parameters, readiness of the organization, and indicators of ML capability measured on a seven-point Likert scale. Several supervised ML models, Linear Regression, Random Forest, Gradient Boosting, and Support Vector Regression are used to capture the linear relationship and non-linear relationships between adoption parameters and supply chain outcomes. This study also used some metrics to measure the performance of our model, such as them being MAE (Mean Absolute Error), RMSE (Root Mean Squared Error), and R². The results demonstrate that ensemble models are superior to the baseline linear models. The most predictive results are those of the Gradient Boosting model with R² = 0.85 (supply chain transparency) and R² = 0.87 (operational efficiency),  as well its lowest MAE = 0.28 (supply chain transparency) and a MAE=0.26 (operational efficiency). Explainable ML model analysis suggests that factors data quality (SHAP = 0.42), traceability capability (0.39), and trust among partners (0.37) are the most important transparency drivers, whereas interoperability (0.41) and machine learning capability (0.45) matter for efficiency. Cost of intervention demonstrates a negative dose-response effect. It is found that combining blockchain-enabled data infrastructures with machine learning analytics leads to a more transparent and efficient US supply chain. These results offer practical implications for managers and policymakers who wish to create smart, transparent, data-driven supply chain systems.