STMicroelectronics, a global semiconductor company serving a wide range of electronics applications, has announced a major step forward in advancing the global development and adoption of physical artificial intelligence (AI) systems. This initiative focuses on enabling next-generation robotics, including humanoid, industrial, service, and healthcare applications, through closer collaboration with NVIDIA.
As part of this effort, STMicroelectronics is aligning its extensive portfolio of advanced robotics components with the NVIDIA Holoscan Sensor Bridge (HSB). This integration is designed to simplify how developers connect sensors and actuators, enabling more efficient data acquisition, synchronization, and processing. At the same time, both companies are incorporating highly detailed simulation models of ST components into NVIDIA’s Isaac Sim platform, strengthening the connection between virtual testing environments and real-world deployment.
Among the first available outcomes of this collaboration is the integration of Leopard Imaging’s depth camera—built using ST technology—with the Holoscan Sensor Bridge. In addition, a high-precision inertial measurement unit (IMU) model developed by ST is now accessible within the Isaac Sim ecosystem, giving developers tools to simulate motion and sensing with improved accuracy.
The collaboration aims to simplify the robotics development process from end to end. By offering pre-integrated solutions that combine STM32 microcontrollers, advanced sensing technologies such as IMUs, imaging systems, and time-of-flight sensors, along with motor control systems, developers can significantly reduce integration complexity. This is particularly valuable for humanoid robotics, where coordination between multiple subsystems is critical.
One of the key challenges in robotics development today is bridging the gap between simulation and real-world performance. High-quality simulations require significant computing resources and carefully tuned parameters. Inaccurate modeling or poorly chosen simulation variables can lead to inefficient training or unreliable results once systems are deployed.
To address this, STMicroelectronics and NVIDIA are working to deliver hardware-calibrated models based on real-world data captured from ST components. These models are optimized for use within NVIDIA Isaac Sim and are expected to include a growing range of devices such as time-of-flight sensors, actuators, and other integrated circuits. By reflecting realistic device behavior, these models allow developers to train AI systems under conditions that closely resemble actual operating environments.
This approach is expected to shorten development cycles, reduce costs, and improve the reliability of robotic systems. Developers will be able to test and refine their designs more effectively before moving to physical prototypes, minimizing trial-and-error during deployment.
The Holoscan Sensor Bridge also plays a central role in standardizing how sensor data is collected and processed. By creating a unified framework for integrating multiple sensors and actuators, it supports the development of scalable AI systems capable of handling complex real-world tasks.
STMicroelectronics brings to this collaboration a well-established ecosystem and deep experience in robotics technologies. Combined with NVIDIA’s expertise in AI computing and simulation platforms, the partnership is positioned to accelerate innovation across industries adopting physical AI.
With robotics becoming increasingly important in sectors ranging from manufacturing to healthcare, the ability to develop systems more quickly and reliably is a critical advantage. Through this collaboration, both companies aim to provide developers with the tools needed to build intelligent machines that can operate safely and efficiently in real-world environments.
