NVIDIA, TI Partner on Physical AI
NVIDIA and Texas Instruments are deepening their collaboration to accelerate the development of physical AI systems like robots and edge devices. The partnership aims to tightly integrate TI's advanced sensing and power electronics with NVIDIA's AI compute platforms like Jetson. This hardware-software co-design is critical for enabling real-time perception and control on power-constrained devices.
This collaboration specifically targets the challenges of real-world humanoid robot deployment by integrating TI's mmWave radar technology with the NVIDIA Jetson Thor robotics computer. The goal is to create a sensor fusion solution that provides low-latency 3D perception, which is critical for robots operating safely in unpredictable environments alongside humans. The partnership leverages NVIDIA's Holoscan Sensor Bridge to connect TI’s IWR6243 mmWave radar sensor to the Jetson Thor via ethernet. This setup is designed to improve a robot's object detection and localization by fusing camera data with radar data, reducing false positives and enabling navigation in difficult conditions like low light, glare, or dust. This move is part of a broader trend where the global AI in robotics market, valued at $22 billion, is seeing rapid growth driven by automation needs in manufacturing, healthcare, and logistics. NVIDIA's Isaac platform, which includes the Jetson line, provides a comprehensive ecosystem for developing, simulating, and deploying AI robots, from software SDKs to virtual simulation environments. For its part, Texas Instruments is a major player in embedded processors for industrial robotics, providing systems-on-chip (SoCs) that combine CPUs, memory, and hardware accelerators. Their components are crucial for real-time motor control, power management, and processing sensor data at the edge. This hardware integration addresses a key safety gap that has limited the deployment of humanoid robots in commercial settings like offices, hospitals, and retail spaces. While vision systems have advanced, they can be unreliable with transparent or reflective surfaces, a problem that radar can mitigate. The demand for engineers with skills in embedded systems, robotics, and AI is surging across sectors like aerospace, defense, and industrial automation. Roles for robotics engineers in the U.S. can command average annual salaries approaching $119,000, with experienced professionals earning more. This collaboration will be showcased at NVIDIA's GTC conference from March 16-19, 2026, in San Jose, demonstrating the system's ability to move from virtual development to production-ready, safety-compliant robots. The partnership extends beyond robotics, as the two companies are also working on 800V high-voltage power systems for the growing energy demands of AI data centers.
