In light of significant advances in the field, the challenges of predicting ball bearing failures persist due to the inherent complexity and dynamic behavior of ball bearing systems. This chapter introduces a state-of-the-art run-to-fail (RTF) testbench for ball bearings, where advanced control and monitoring technologies are employed. The primary objective behind this RTF testbench architecture is to establish baseline data for ball bearing failures, serving as a validation platform for emerging Artificial Intelligence-based predictive maintenance technologies. Within this ball bearing testbench, Aingura IIoT Company conducts accelerated degradation tests under genuine operational conditions, meticulously controlled by numeric control actuators. Simultaneously, the testbench employs high-speed data acquisition and online pre- and processing Machine Learning algorithms to extract actionable insights, such as remaining useful life (RUL). The overarching goal of the testbench is to faithfully replicate real-world operational scenarios by subjecting ball bearings to diverse mechanical forces, vibrations, and thermal influences, thereby generating valuable insights into operating conditions and failure modes, all derived from the collected data. Additionally, this chapter addresses the challenges associated with testbench development and operation, providing a comprehensive explanation of its constituent subsystems. This chapter also presents RTF test examples with real operational conditions and results conducted on the testbench.