New Sensor Frontiers From Jestson Thor Robotic Platform
The availability of Nvidia’s Jetson Thor redefines what’s possible for humanoids and autonomous mobile robots (AMRs) by fusing physical intelligence and real-time reasoning through robotics foundation models.
The new compute platform for robots comprises a Blackwell GPU, transformer engine, multi-instance GPU (MIG), a 14-core Arm Neoverse V3AE CPU, and up to 128 GB of LPDDR5X memory.
The platform’s high-throughput I/O, including 4×25 GbE, provides the bandwidth needed to fuse dense multimodal sensing in real-time. Nvidia’s Jetson Thor, a Blackwell-powered robotics computer, is the first platform to run robotics foundation models at scale; these models compress decades of challenges into perception-rich humanoids capable of dexterous, human-speed manipulation.
However, the real breakthrough, according to Paul Golding, VP of Edge AI at Analog Devices Inc., is reasoning that integrates multimodal inputs to plan, adapt, and act in real-time. “For the first time, robots can understand complex tasks.”
Next, Sebastian Scherer, an associate research professor at Carnegie Mellon’s Robotics Institute, points out that there has been a big disconnect between computer vision and robotics because computer vision workloads were too slow for real-time decision-making. With Jetson Thor, models and computing have become fast enough, allowing robots to handle much more nuanced tasks.
Jetson’s ecosystem—which encompasses Holoscan Sensor Bridge and Isaac Sim—supports a variety of application requirements, high-speed industrial automation protocols, and sensor interfaces. Holoscan Sensor Bridge is a sensor-over-Ethernet technology designed to enable real-time data streaming and simplify high-speed sensor fusion and actuator integration on Nvidia’s edge AI platforms.
Sensor and actuator companies—including ADI, e-con Systems, Infineon, Leopard Imaging, and RealSense—are using this platform to connect sensor data from cameras, radar, and lidar, and send it directly to GPU memory on Jetson Thor with ultra-low latency.
Isaac Sim is an open-source reference application that enables developers to simulate and test AI-driven robotics solutions in physically based virtual environments. ADI is embedding robotics foundation models into its development stack so that its hardware behaves in Nvidia Isaac Sim as it will in the real world.
“Our goal is to build the most physically accurate robotics content in Nvidia Isaac Sim, enabling teams to iterate at simulation speed and then scale seamlessly to real systems with ADI hardware and Nvidia Jetson Thor,” said ADI’s Golding.
In this case, physical intelligence fuses sensing, actuation, policy learning, and reasoning to enable robots to execute precise industrial tasks. At the same time, however, it demands high-fidelity edge sensing, energy-efficient and functionally safe power, and deterministic connectivity to central compute.
That opens the doors for new opportunities for analog chip firms with expertise in edge sensing, precision motion control, power integrity, and deterministic connectivity. When combined with Jetson Thor’s compute capabilities, these analog building blocks can accelerate the development of humanoids and AMRs. |
