Modern enterprises operate in a data-driven world where sensitive data (e.g., customer information, intellectual property, financial records) are both a critical asset and a significant liability. Although leveraging these data can unlock innovation, improve decision-making, and personalize customer experiences, mishandling it risks breaches, regulatory penalties, and loss of trust. How to maximize data utility while ensuring privacy and security is a core challenge in enterprise AI. Federated learning (FL) has emerged as a groundbreaking paradigm in the realm, effectively addressing critical data privacy and security concerns while facilitating the collaborative development of robust AI models. Its versatility and effectiveness have been demonstrated across various industries, including healthcare, finance, advertising, and personalized recommendations, where it has been successfully employed to address data heterogeneity and privacy concerns. This chapter aims to provide a comprehensive introduction to the fundamental concepts of FL, encompassing categorization, formulation, and challenges. We will also utilize a federated recommendation system as a case study to illustrate the application of FL in real-world scenarios with diverse settings. Furthermore, we will explore the role of federated learning in refining and adapting large foundation models to meet industrial requirements. By grasping the fundamental concepts and applications of FL, readers can gain a thorough understanding of this approach and select the most suitable techniques for their specific use cases.