The Japan Science and Technology Agency (JST) is advancing a bold vision for the future with the Moonshot research program, a national initiative aiming to integrate AI-powered, autonomously learning robots into everyday life by 2050. Central to this mission is the development of advanced caregiving robots designed to support Japan’s rapidly aging population. A key component of the program is the AI-Driven Robot for Embrace and Care, or AIREC, a family of humanoid robots developed with the help of NVIDIA’s AI and robotics technologies. The larger, more mobile version, Dry-AIREC, is equipped with two NVIDIA GPUs, enabling high-performance AI processing for complex tasks. AIREC-Basic, used primarily for data collection to train a motion foundation model, runs on three NVIDIA Jetson Orin NX modules, providing powerful edge AI capabilities. The project also leverages NVIDIA Isaac Sim, an open-source robotic simulation platform, to train robots in virtual environments before real-world deployment. This has been critical in teaching the robots to perform delicate actions like estimating forces between objects during physical interactions. Tetsuya Ogata, professor and director of the Institute for AI and Robotics at Waseda University, noted that just five years ago, such applications seemed unattainable. “Now, the atmosphere surrounding this technology has changed, so we can seriously think about this kind of application,” he said, highlighting the transformative role of generative AI and modern hardware. One of the most challenging tasks being developed is repositioning elderly patients in bed—a crucial action to prevent bedsores and support hygiene. Misa Matsumura, a bioengineering master’s student at the University of Tokyo, is leading research in this area. Her team used laptops powered by NVIDIA RTX GPUs to run 3D posture estimation, trajectory planning, and force prediction models. Using fisheye and depth cameras, the robot analyzes patient movement and determines the optimal repositioning path based on data from skilled human caregivers. The system also calculates the right amount of force to apply, ensuring patient comfort and safety. Initial tests were conducted with mannequins, and the project has now moved to human trials. Matsumura continues refining the robot’s performance to make it more responsive and adaptive. Etsuko Kobayashi, a professor of bioengineering at the University of Tokyo and Matsumura’s advisor, joined the project not only for its technical promise but also for personal reasons. “My mother is growing older, and that experience has given me an appreciation for the importance of personal care,” she said. “I found that my background in medical robotics can be meaningfully extended to care robotics, contributing to safe and reliable systems for human-centered applications.” The Moonshot team is set to present its progress at the 2026 International Symposium on System Integration in January. With NVIDIA’s AI infrastructure enabling rapid development and simulation, the vision of robots assisting in elder care is moving from science fiction toward reality.