PSYONIC & NVIDIA Officially Announce Collaboration at NVIDIA GTC 2026
PSYONIC / NVIDIA GTC PRESS RELEASE - March 16, 2026
PSYONIC Ability Hand Becomes One of the First Commercial Dexterous Hand Officially Integrated into NVIDIA Isaac Lab.
PSYONIC also announces collaboration with NVIDIA to introduce human to robot “real-to-real" transfer to accelerate robotic manipulation learning.
PSYONIC today announced that the PSYONIC Ability Hand is now available as a native asset within NVIDIA Isaac Lab, an open-source framework for robot learning, becoming the first commercially deployed dexterous hand integrated directly into the platform.
Through its collaboration with NVIDIA, PSYONIC is also introducing human to robot “real-to-real" transfer, a novel approach for capturing high-fidelity human dexterous interaction data to accelerate robotic manipulation learning. By bringing the Ability Hand into Isaac Lab as a native asset, researchers and developers can simulate, train, and validate policies on a high-performance, sensorized bionic hand already deployed in real-world human applications. This integration establishes a unified pipeline connecting assistive bionics and embodied AI research.
PSYONIC builds advanced, accessible bionic systems designed to supercharge both humans and robots. The Ability Hand is the fastest dexterous hand on the market and one of the first to provide touch feedback to users. Its combination of speed, durability, and integrated tactile sensing makes it not only a leading dexterous hand for human users, but also a powerful data-generation platform for robotics and AI research.
As part of this announcement, PSYONIC is showcasing a video demonstration of real-to-real transfer centered on a precise pipetting task. In the demonstration, a human user wearing the Ability Hand finely squeezes and fills a pipette with food coloring, moves it above another glass, and precisely dispenses drops into a glass of water. The same Ability Hand is then shown connected to multiple robots - an industrial robot arm, a mobile robot dog, and a humanoid robot - each performing the same manipulation task.
The demonstration illustrates a core principle behind real-to-real transfer: regardless of the robot, the highest quality dexterous manipulation policies will originate from human data. Rather than relying solely on synthetic simulation data, real-to-real transfer captures true, authentic human manipulation behavior in uncontrolled environments from the same hand being used on both humans and robots. These high-resolution, human-guided datasets complement simulation workflows in Isaac Lab and revolutionize the development of more robust and generalizable manipulation policies.
The native Isaac Lab integration enables a closed-loop workflow: simulation and policy development in Isaac Lab, deployment on physical hardware, human-guided data capture in the real world, and subsequent robotic training informed by that data. This unified pipeline supports scalable improvements in dexterity across assistive bionic systems and diverse robotic platforms.
“Robots will learn the real world from humans," said Dr. Aadeel Akhtar, Founder and CEO of PSYONIC. “We are excited to officially collaborate with NVIDIA to enable real-to-real transfer by capturing high quality human dexterous interaction data that can be used to train robots to perform the same real world tasks. The lack of high quality manipulation data is one of the biggest challenges in robotics, and this collaboration is about building that foundation, improving dexterity for both humans and robots."
PSYONIC's collaboration with NVIDIA reflects a shared commitment to advancing embodied intelligence at the intersection of robotics and AI. As autonomous systems require increasingly sophisticated interaction with unstructured environments, access to high-quality, physically grounded manipulation data becomes critical. The Ability Hand's commercial deployment and integrated sensing architecture uniquely position it as a bridge between human capability and robotic learning. Dr. Aadeel Akhtar will attend NVIDIA GTC to discuss the Isaac Lab integration and the broader implications of real-to-real transfer for embodied AI. He will be joined by PSYONIC's Creative Marketing Manager, Dale DiMassi, who is also an Ability Hand user, offering his perspective on the human experience that informs this approach to robotic learning.
About PSYONIC
Founded in 2015 and based in San Diego, PSYONIC develops advanced, accessible bionic systems designed to supercharge both humans and robots. The company builds advanced bionics that combine speed, strength, durability, and sensory feedback while advancing research at the intersection of robotics, brain-computer interfaces, and AI. The PSYONIC Ability Hand is the fastest dexterous hand on the market and the first to deliver touch feedback, enabling intuitive physical interaction for human users and generating rich datasets to improve robotic dexterity.
Contact:
PSYONIC, Inc.