AI-Driven Quantum-Resistant Security Framework for Accelerating the Migration of U.S. Cloud and IOT Systems to NIST Post-Quantum Cryptography Standards

Abstract

The accelerating development of quantum computing creates substantial cybersecurity challenges for cloud and Internet of Things (IoT) environments that continue to depend on quantum-vulnerable public-key cryptography. This study developed and quantitatively evaluated an AI-driven quantum-resistant security framework for accelerating the migration of U.S. cloud and IoT systems toward National Institute of Standards and Technology (NIST) post-quantum cryptography (PQC) standards. The final analytical environment included 52 organizations, 164 cloud environments, 18,764 IoT devices, and 139,684 retained cryptographic assets, with 98.7% data completeness and 97.9% inventory-validation accuracy. Among retained assets, 62.1% contained quantum-vulnerable dependencies, including 34.6% associated with RSA, 20.0% with elliptic-curve cryptography, and 7.5% with Diffie–Hellman, while 63.2% of certificates remained quantum vulnerable. AI-assisted cryptographic discovery achieved 97.4% asset coverage and 97.9% vulnerability-detection sensitivity while reducing assessment time by 87.8%. XGBoost achieved the strongest quantum-risk classification performance, with 97.6% accuracy, 97.5% F1-score, 0.994 ROC-AUC, and 0.992 PR-AUC. Experimental evaluation of ML-KEM, ML-DSA, and SLH-DSA showed that pure PQC increased cloud secure-session latency by 23.8% and reduced throughput by 10.0%, whereas hybrid configurations increased latency by 39.5% and reduced throughput by 15.6%. AI-assisted migration increased implementation success from 84.6% to 97.1%, reduced remediation time by 60.2%, increased migration throughput by 151.1%, and reduced unresolved quantum-vulnerable exposure from 15.4% to 2.9%. High-security-maturity organizations achieved 98.8% implementation success compared with 87.8% among low-maturity organizations. Validation confirmed strong generalizability, with XGBoost achieving 97.1% cross-validation accuracy, 96.9% independent holdout accuracy, and 95.9% temporal-validation accuracy. Overall, the findings demonstrated that integrating AI-assisted cryptographic discovery, quantum-risk classification, migration prioritization, PQC performance assessment, cryptographic agility, interoperability testing, and organizational readiness provided a robust quantitative framework for strengthening post-quantum security across heterogeneous U.S. cloud and IoT infrastructures.