Exploring physical AI development across industries

This initiative aims to promote the development of a collaborative control platform that ensures sovereignty by bridging the digital and physical worlds and integrating NVIDIA’s open physical AI technologies. By accelerating the societal implementation of physical AI across various industrial sectors – including manufacturing, logistics, and healthcare – Fujitsu seeks to realise a society where humans and robots coexist and collaborate, and strengthen Japan’s industrial competitiveness.

In recent years, various industrial sectors, particularly manufacturing, have faced growing challenges such as labour shortages due to Japan’s declining birthrate and aging population, a decline in the number of skilled technicians, and intensifying global competition. To resolve these challenges and achieve sustainable growth, the promotion of digital transformation (DX) is essential. In particular, expectations are rising for physical AI, where AI recognises and analyses real-world information and executes it as physical actions.

Physical AI enables the automation of tasks, improved productivity, stable quality, and the creation of new services by allowing robots and various pieces of equipment to assess situations and autonomously determine and execute optimal actions. However, realising this requires advanced robot control technology and an AI infrastructure that leverages high-quality on-site data, as well as a collaborative control platform that integrates these elements to bridge the digital and physical worlds. Yet, there are limits to what a single company can achieve in terms of development and widespread adoption.

Given this situation, as a technology company, Fujitsu will promote the societal implementation of physical AI through collaboration with leading robotics companies—FANUC, Yaskawa Electric, and Kawasaki Heavy Industries, while simultaneously working to standardise and open up the collaborative control platform.

Overview

Social implementation utilising physical AI

These explorations will cover the social implementation of physical AI in the following industrial sectors with plans to expand into other sectors in the future in place.

• Factory solutions: By optimising the planning of overall factory production activities,taking into account factors affecting production fluctuations and on-site conditions, and enabling autonomous on-site adaptation, Fujitsu will help the manufacturing industry achieve further productivity gains and flexibility.
• Solutions for retail and logistics: By automating material handling operations based on logistics plans that incorporate real-time sales and inventory data, Fujitsu will achieve labour-saving and automation in logistics.
• Healthcare solutions: Based on optimised plans triggered by instructions from hospital operational systems, robots autonomously execute tasks to automate the in-hospital transport of pharmaceuticals and specimens, as well as outpatient reception and guidance services.

Fujitsu will develop software platforms and hardware interfaces that serve as a common foundation for physical AI by first gaining a deep understanding the cutting-edge technologies – such as AI, robotics, control systems, simulation, and data analysis – held by each company. These technologies will facilitate collaboration among various robots and equipment, and help to realise more advanced autonomous control systems. At the same time, as the scope of robot applications expands and collaboration with other equipment increases, risks such as cyberattacks, system-wide downtime or malfunctions, and leaks of confidential information also rise. Therefore, Fujitsu will develop a sovereign collaborative control infrastructure which will be provided as an open platform for participating companies and research institutions, thereby promoting the implementation of physical AI across the entire industry.

Collaboration with NVIDIA

Through this initiative, Fujitsu will lead business discussions with the companies and leverage the AI, world model, simulation, and robotics technologies underpinning NVIDIA’s physical AI platform. By doing so, it will further enhance the sovereign collaborative control platform and accelerate the social implementation of physical AI in industrial domains.

• Fujitsu will leverage the NVIDIA Cosmos global foundation model in its socio-physical simulations to enhance its ability to understand and predict events in real-world environments. Furthermore, it will accelerate the development and real-world implementation of physical AI solutions across diverse sectors, including manufacturing, logistics, and healthcare.
• Fujitsu will leverage libraries such as NVIDIA Omniverse, the NVIDIA Isaac open platform, and the Newton physics engine to streamline Sim2Real as well as robot learning, verification, and optimisation.

Future plans

Starting with business discussions with these companies, Fujitsu will formulate a roadmap for concrete technology development and business expansion. As the robotics market expands, Fujitsu is confident that establishing a framework to provide an AI infrastructure, leveraging Japan’s world-class robot control technology and high-quality on-site data, will be a crucial step toward Japan leading the global robotics market. Through the societal implementation of physical AI, Fujitsu will contribute to the realisation of a safe and prosperous society where humans and robots coexist and collaborate, as well as to the strengthening of Japan’s industrial competitiveness.

global.fujitsu