<p>The capacity of machine learning (ML) to make accurate predictions based on data is transforming the process of discovering, designing, and implementing advanced materials, as it allows breaking the constraints of existing experimental and computational methods. It revolutionizes the understanding, design, and deployment of advanced materials. The present review is a critical overview of the state of ML in nanoscience, including property prediction, optimization of synthesis, high-throughput screening, inverse design, and real-time characterization. It focuses on the application of supervised, unsupervised, semi-supervised, and reinforcement learning approaches within various classes and functions of nanomaterials. Moreover, we discuss the use of generative models and hybrid frameworks to support autonomous discovery of materials as well as the discovery of sustainable nanotechnologies. By integrating current approaches with future trends, this paper summarizes the strong strategic promotion of ML, which enables faster innovation, eliminates resource-intensive tasks, and increases reproducibility. The first originality of the review lies in its cross-domain approach, which integrates the technological, economic, and environmental aspects of ML-driven nanomaterials research. This synthesis serves as both a roadmap and an action agenda for the design of intelligent materials, with applications in the development of electronics, energy industries, catalysis, and biomedical fields, in the context of information-intensive science.</p> Graphical Abstract <p></p>