Cloud and Enterprise Networks are the foundation of today's digital age, facilitating frictionless communication and service delivery. Cloud and Enterprise Network attacks increasingly depend on trusted protocols, and DNS Data Exfiltration Attacks in Cloud and Enterprise Networks have evolved as a devious and powerful means to evade classic defences. Detecting DNS Data Exfiltration Attacks in Cloud and Enterprise Networks is therefore a pressing challenge that requires efficient and accurate solutions. This study investigates Machine Learning Models for Detecting DNS Data Exfiltration Attacks in Cloud and Enterprise Networks, focusing on lightweight approaches such as Random Forest, Decision Tree, Multi-Layer Perceptron, Logistic Regression, and Gaussian Naïve Bayes. Both Random Forest and Decision Tree achieved perfect evaluation scores (100%) across standard metrics, but closer inspection of confusion matrices revealed Random Forest as the superior model, misclassifying only two malicious instances while generating no false positives. The significance of this research lies in demonstrating that lightweight models, particularly Random Forest, can provide highly accurate, resource-efficient, and practical real-time protection against DNS exfiltration threats, ensuring the resilience of cloud and enterprise infrastructures.