SoftBank is stepping up its game in robotics and AI infrastructure. The company is in talks to invest about $800 million in Agile Robots, a Munich-based startup. Agile Robots makes robot arms, warehouse machines, and humanoids. This would be one of SoftBank’s biggest bets yet on physical robots.

The deal would see SoftBank contributing over $300 million of the total round. SoftBank first backed Agile Robots in 2021. That round helped turn the company into Germany’s first robotics unicorn. Now, SoftBank looks ready to deepen its stake in the startup.

Agile Robots was created by researchers from Germany’s space agency in 2018. It now employs more than 3,200 people across Germany, China, and India. The company builds both the software and hardware for its machines. Investors like this vertical approach since it controls the whole robotics stack.

Investment in physical AI and robotics is booming. In 2025, global funding for robotics and physical AI hit $27.6 billion. That was more than twice the amount from the previous year. Investors see a big opportunity in applying AI to machines that work in factories, warehouses, and logistics centers.

SoftBank’s Vision for AI and Robotics

Masayoshi Son, SoftBank’s CEO, has been a big believer in robotics for years. SoftBank launched the Pepper humanoid robot back in 2014 but stopped production in 2020. Since then, the company has shifted focus to industrial robotics and AI infrastructure.

Last year, SoftBank agreed to buy the industrial robotics arm of Swiss company ABB for $5.4 billion. It is also planning to create a new AI and robotics company called Roze. Son’s interest in European robotics is part of a bigger strategy, including a €75 billion plan to build AI data centers in France.

Son sees AI as a technological shift even bigger than the internet boom. He said AI could be 50 times more impactful than the dot-com era. This belief drives SoftBank’s massive investments in AI infrastructure across Europe and the US.

Building AI Infrastructure in Europe

SoftBank’s €75 billion investment in France aims to build 3.1 gigawatts of AI data center capacity by 2031. These data centers will be located in regions like Dunkirk, Bosquel, and Bouchain. The project will mostly rely on project financing and support from expected customer demand.

This effort is part of SoftBank’s global AI infrastructure push, which also includes the Stargate project in the US. Stargate is a collaboration with OpenAI to build large-scale AI computing systems. SoftBank’s portfolio includes a majority stake in Arm, the British chip designer, which is crucial for AI hardware.

By controlling key parts of the AI supply chain, SoftBank aims to benefit no matter which AI applications succeed. Son’s focus is on building the hardware and data infrastructure that will power next-generation AI, including Artificial General Intelligence (AGI).

AGI refers to AI systems that can perform any intellectual task a human can. Son calls AGI the next trillion-dollar opportunity. He believes it will transform industries and society over the next 50 to 100 years. This long-term view shapes SoftBank’s investment decisions.

Japan’s Domestic AI Venture and Industry Backing

In Japan, SoftBank is launching a new AI venture focused on “physical AI.” The goal is to develop AI that controls machines and robots autonomously. About 30 Japanese companies, including those from chemicals, robotics, automotive, and electronics sectors, are considering backing this project.

This collaboration aims to train AI systems using manufacturing data from materials and machinery. It reflects SoftBank’s broader strategy to strengthen domestic AI capabilities and create smarter industrial robots.

The move aligns with SoftBank’s global investments. By combining European robotics startups, AI data centers in France, and domestic AI ventures in Japan, SoftBank is covering all bases. The company is positioning itself to lead the next wave of AI-powered physical technology.

Robotics and AI hardware are expensive and complex, but SoftBank’s strategy focuses on owning the foundational technology. Instead of betting on individual software startups, it invests in the chips, data centers, and machines themselves. This approach could pay off as AI spreads into the physical world.