Humanoid robots are moving closer to real-world deployment – and their progress depends on physical intelligence and real-time reasoning. With the recent announcement of general availability of NVIDIA Jetson Thor, Analog Devices is further accelerating the development of humanoids and autonomous mobile robots (AMRs).
Jetson Thor redefines what’s possible for robotics. With a NVIDIA Blackwell GPU, transformer engine, Multi-Instance GPU (MIG), a 14-core Arm Neoverse V3AE CPU, and up to 128GB of LPDDR5X memory, it delivers 2070 FP4 TFLOPS server-class AI compute in a mobile power envelope. Its high-throughput I/O, including 4×25GbE, provides the bandwidth needed to fuse dense multimodal sensing in real time.
This capability makes NVIDIA Jetson Thor the first platform to run robotics foundation models at scale, from vision-language to vision-language-action models, enabling robots to move beyond perception into reasoning and physically intelligent behaviour. That aligns directly with ADI’s R&D focus: sensing, perception, control and connectivity that makes such reasoning actionable in the real world with high physical accuracy.
Robotics foundation models compress decades of challenges into perception-rich humanoids capable of dexterous, human-speed manipulation. But their real breakthrough is in reasoning: integrating multimodal inputs to plan, adapt and act in real time.
ADI are embedding robotics foundation models into the ADI development stack, closing the Sim2Real gap so the hardware behaves in NVIDIA Isaac Sim as it will in the real world. The goal will be 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.
Physical intelligence fuses sensing, actuation and policy learning and reasoning so robots can execute precise industrial tasks. It demands high-fidelity edge sensing, energy efficient and functionally safe power, deterministic connectivity to central compute, and a digital twin that closes the Sim2Real loop.
What ADI brings to humanoids
● High-fidelity edge sensing for contact-rich manipulation: ADI provide novel multimodal tactile sensing in development, plus ToF depth, high-accuracy IMUs, joint encoders, and multi-axis force/torque sensors to capture contact and proprioception with precision.
● Precision motion and functionally safe power control: ADI offers drivers and control for current, position, and torque – along with advanced multi-turn magnetic sensors – to deliver accurate, energy-efficient, and safe actuation.
● Deterministic connectivity to central compute: ADI’s capabilities include time-synchronised data paths, integrated with Holoscan via custom operators optimised for ADI’s data-fabric stack, to enable bounded-latency ingest of dense sensor and perception flows into NVIDIA Jetson Thor.
● Simulation and digital-twin fidelity: ADI’s sensor models and parameterised device behaviour for NVIDIA Isaac Sim/Omniverse matches the ADI hardware, improving policy transfer and task completion from simulation to real systems.
How ADI’s robotics stack maps to Jetson Thor
● Holoscan Sensor Bridge delivers deterministic ingest with NVIDIA JetPack 7, streaming synchronised ADI sensor/actuator data into NVIDIA Jetson Thor’s GPU/CPU with bounded latency using Holoscan operators optimised for ADI’s data stack.
● 4×25GbE interconnect provides high-throughput, time-aligned fusion across hands, arms, torso, and perception nodes, with ADI connectivity keeping the sense–reason–act loop synchronised and low latency.
● Thor’s 2070 FP4 TFLOPS capacity supports foundation models like NVIDIA Isaac GR00T alongside VLM/LLM reasoning, while ADI tactile, ToF depth, IMU, and encoder inputs enrich training and runtime policies, bringing physical accuracy to real-time reasoning.
● MIG-based workload partitioning enables isolated GPU slices for locomotion, grasp planning, perception, and VLA policies, simplifying functional decomposition.
“With NVIDIA Jetson Thor as the brain and ADI’s high-fidelity sensing, signal-chain fidelity and deterministic connectivity as the nervous system, we take robots from NVIDIA Isaac Sim to the factory floor with physical accuracy – faster.” — Paul Golding, VP of Edge AI, ADI
https://www.nvidia.com
