Addverb Accelerates Physical AI for Industrial Robotics with Newton and NVIDIA Isaac

Addverb, a global robotics and warehouse automation company, today announced the expansion of its end-to-end robotics development workflow using NVIDIA AI, simulation, and edge computing platforms. This enhanced workflow is being deployed across Addverb’s Trakr quadruped robot and Elixis-W wheeled humanoid, enabling faster design, training, testing, and large-scale deployment of next-generation industrial robots.

As part of this initiative, Addverb is leveraging NVIDIA Omniverse’s libraries and open NVIDIA Cosmos World Foundation Models to build high-fidelity digital twins of real-world warehouses and industrial environments. These physically based virtual environments enable large-scale testing and optimisation using synthetic data, helping validate robot behaviour earlier in the development cycle and significantly reducing time to deployment.

Commenting on the expansion, Mr. Sangeet Kumar, Co-founder & CEO, Addverb, said, “By combining digital twins, NVIDIA GPU-accelerated robot learning, and edge AI through our collaboration with NVIDIA, we are strengthening our ability to validate robotic systems faster, improve sim-to-real confidence, and reliably scale deployments across diverse industrial environments. This collaboration has been instrumental in accelerating innovation while ensuring our robots perform consistently and safely in real-world shopfloor conditions.”

To strengthen sim-to-real transfer, Addverb is adopting NVIDIA Isaac Lab, a GPU-accelerated, open-source, robot learning framework built on NVIDIA Isaac Sim. Isaac Lab enables scalable training and evaluation of robot policies while supporting reinforcement learning workflows within a unified development stack.

Addverb is also evaluating simulation workflows enabled by Newton, an open-source physics engine co-developed by NVIDIA, Google DeepMind, and Disney Research, and managed by the Linux Foundation . Additionally, Adverb is exploring server learning and edge deployment on NVIDIA® Jetson Thor for Vision Language Action [VLA] models.

For deployment in industrial environments, Addverb’s robots are powered by NVIDIA Jetson Orin NX, with NVIDIA® TensorRT enabling low-latency, high-performance inference at the edge. This supports real-time perception, navigation, and decision-making across complex warehouse and industrial operations.

“Advances in robotics require a seamless integration of AI, simulation, and edge computing,” said Mr. Amit Goel, Head of Strategic Robotics Partnerships at NVIDIA. “Our collaboration with Addverb reflects NVIDIA’s commitment to empowering partners with the tools and platforms needed to accelerate innovation, improve reliability, and bring scalable automation to industrial environments worldwide.”

This expanded workflow reinforces Addverb’s focus on building production-grade Physical AI systems that deliver intelligent, scalable automation for warehouses and industrial facilities worldwide.

Addverb’s Physical AI robots, powered by this enhanced NVIDIA-enabled workflow, are currently on display till February 20, 2026, at Bharat Mandapam in New Delhi as part of the India Impact AI Summit 2026, demonstrating intelligent, production-ready automation for real-world industrial environments.