Texas Instruments (TI) and Nvidia are collaborating on new semiconductors that will accelerate the path from simulation to the deployment of humanoid robots.
Under the deal, TI’s real-time motor control, sensing, radar and power technologies will be combined with Nvidia’s advanced robotics compute system, Ethernet-based sensing and simulation technologies.
This will allow robotics vendors to validate perception, actuation and safety earlier and more accurately, the companies said. As part of the collaboration, TI designed a sensor fusion solution by integrating its millimeter wave (mmWave) radar technology with Nvidia’s Jetson Thor using Nvidia Holoscan sensor bridge.
This will enable humanoid robots to have better:
- Low latency
- 3D perception
- Safety awareness
Nvidia and TI will demonstrate the semiconductor solution at Nvidia GTC during March 16-19, 2026, in San Jose, California.
"The next generation of physical AI (artificial intelligence) requires more than just advanced compute – it demands seamless integration between sensing, control, power and safety systems," said Giovanni Campanella, GM of industrial automation and robotics at TI.
Campanella said that TI will bridge the gap between AI compute and real-world applications to accelerate the evolution from prototypes to commercially viable humanoid robots that will work alongside humans.
Human-like perception
TI said by fusing camera and radar data, the semiconductor offering improves:
- Object detection
- Localization
- Tracking
Additionally, it reduces false positives for real-time decision-making in humanoid robots.
This human-like perception works in difficult conditions like low light and bright glare and even fog and dust. It also works indoors and outdoors. This ability to see in challenging conditions is a gap that has eluded the deployment of humanoid robots in real-world deployments, TI said.
TI gave an example of how cameras do not reliability detect glass doors or reflective surfaces but radar provides consistent detection of transparent objects. This allows robots to navigate areas with various surfaces like offices, hospitals and retail shops.
