Industry News
2025-12-09
NVIDIA Research Center Promotes Physics AI development

For a long time, medical and industrial enterprises have been promoting medical imaging, surgical robots, and industry AI  Automation and other applications face many limitations due to the lack of speed and capability of edge systems to process real-time data, or the inability to safely run multiple generative systems without affecting performance AI  Model.

IGX Thor  By providing powerful and reliable solutions tailored for the industrial and medical fields AI  Computational power has overcome these challenges. This platform has real-time intelligent processing, seamless data connection, and built-in security protection functions.

This platform is equipped with two models NVIDIA Blackwell GPU—— integration GPU (iGPU)  And independence GPU (dGPU), 5581 can be provided FP4 TFLOPS  of AI  Computational power and support for 400 GbE  connect. and NVIDIA IGX Orin  comparison, IGX Thor  in iGPU  on AI  The computing power can be increased by 8 times dGPU  Up to 2.5 times higher, and at the same time, the connectivity performance is improved by 2 times, allowing seamless operation of large language models and visual language models at the edge.

This industrial grade platform provides 10-year product lifecycle support and NVIDIA AI  Long term technical support for the software stack ensures both enterprise and physical support AI  Applications can maintain acceleration and safe operation even in the most challenging environments. IGX Thor  Can support running NVIDIA AI Enterprise  Software stack, including physical acceleration from cloud to edge devices AI  Application Development NVIDIA NIM  Microservices, used for robot development NVIDIA Isaac、 Used for visual purposes AI  of NVIDIA Metropolis  And for sensor processing NVIDIA Holoscan。

Besides, NVIDIA IGX Thor  Will also NVIDIA Halos  Integrating elements of a full stack security system into robot development, healthcare, and industry AI  In the system, onboard and infrastructure sensors are utilized to ensure traditional and external safety, achieving reliable human-machine collaboration.