onsemi Targets Physical AI with Automotive Power Tech

Recap

Source: 테크월드, report of Sept. 30, 2026

According to the report, onsemi is leveraging its automotive power and sensing expertise to enter the physical AI market. Sudhir Gopalswamy, head of the Analog/Mixed-Signal and Intelligent Sensing Group, stated at an Investor Day that the company's next growth phase involves increasing semiconductor content and applying automotive innovation to new industrial systems.

The company outlined a silicon carbide (SiC) roadmap that aims to improve device performance by approximately 25% per generation. It plans to extend its roadmap to 1400V SiC to support next-generation 1000V battery architectures. Onsemi also introduced its 'Treo' 65nm analog and mixed-signal platform, which integrates power, sensing, and control functions. The company is expanding this into an Embedded Power Platform (EPP) that targets up to 4x power density and a 15% reduction in power loss compared to traditional inverter designs.

In the sensing sector, onsemi is combining image, ultrasonic, and inductive sensing technologies. It is expanding into indirect time-of-flight (iToF) and short-wave infrared (SWIR) sensing, aiming to improve sensing performance by up to 20x while reducing system costs through integration. The company views physical AI as a natural extension of autonomous driving, noting that humanoid robots and other intelligent machines require high semiconductor content for sensing, power management, and connectivity.

Context

Physical AI refers to AI systems that interact with the physical world, encompassing humanoid robots, drones, and industrial automation. This sector is currently seeing significant investment as companies seek to bridge the gap between digital models and real-world execution. onsemi has long been a leader in silicon carbide technology for electric vehicles, positioning its power management expertise as a transferable asset for the emerging robotics market.

Robot's take

onsemi's strategy suggests a shift from selling discrete components to providing integrated power and sensing solutions for embodied AI. The focus on high-density power delivery and advanced sensing is well-suited for the demanding requirements of humanoid robots. However, the success of this pivot will depend on how well these automotive-grade solutions can be adapted to the specific control and latency needs of robotic platforms. It remains to be seen if the company can effectively compete in a market where software integration is often as critical as the hardware.