The integration of electric motors into various industrial and automotive applications emphasizes the critical necessity for reliable performance and operational efficiency. The advent of advanced digital technologies offers opportunities for predictive maintenance strategies. Digital Twins (DTs), mathematical models simulating a system's physical behavior in real-time, present a transformative approach to enhance real-time monitoring of critical quantities, which is imperative to improve operational efficiency and minimize downtime. In this paper, we explore the feasibility and efficacy of deploying real-time physics-based DTs for condition monitoring in electric motor applications. Particularly, we focus on employing on-the-edge DTs, implemented on low-power onboard microprocessors, ensuring continuous communication with the physical asset for reliable real-time monitoring. The study applies DT technology to a high-voltage high-density Electric Vehicle (EV) motor, assessing its predictive capabilities in a real-world scenario. Results showcase the potential of DTs in revolutionizing condition monitoring, thereby meeting the evolving operational and maintenance requirements of contemporary electric motor systems.