Agile Robots Raises $500M Series F for Physical AI Growth

A robot's working day can supply part of the material used to teach its successor. Agile Robots is organizing its expansion around that connection between industrial deployment and machine learning, bringing a manufacturer of arms, hands and mobile systems further into the business of physical intelligence.

The Munich company closed a $500M Series F announced November 12, 2026, at an $8.5B valuation. SoftBank Group led, joined by NVIDIA, Hitachi Group, including Hitachi and Hitachi Ventures, T.Capital and others. The financing brings total invested capital to $2B and supports US expansion, additional production capacity including Germany, and new markets. For industrial buyers, the interesting development is the combination of machinery, AI research and routes to deployment now gathering around the same supplier.

Industrial experience becomes part of the AI development process

Zhaopeng Chen and Peter Meusel founded Agile Robots in 2018 after working at the German Aerospace Center. Chen, the company's founder and CEO, had already pursued humanoid-hand research and an earlier robotics spin-off with Meusel. That history helps explain a business whose ambitions reach into the detailed mechanics of manipulation as well as the software deciding how a task gets performed.

Agile's industrial portfolio gives those ambitions several places to develop. Robot arms, mobile systems and humanoids encounter different parts of a production process, with different requirements for movement, handling and coordination. A customer may need several kinds of equipment to cooperate before an improvement becomes useful to the operation. The purchasing decision therefore extends beyond the capabilities of one robot shown in isolation.

The company's March research partnership with Google DeepMind makes the learning relationship explicit. The partners combine Gemini Robotics models with Agile hardware, connecting deployment, data collection, training and iteration. In commercial terms, this creates a reason to care about what happens after installation: work performed in an industrial setting can inform the next development cycle. Keeping that cycle productive across varied environments adds another responsibility to the manufacturer's job, alongside building the machine and supporting the customer using it.

Hitachi brings a commercialization role alongside its capital

Hitachi's place in the financing follows a September 29 co-creation agreement with Agile Robots. The companies intend to combine complementary AI robotics technologies and industrial knowledge. Hitachi plans to offer the resulting systems globally through HMAX Industry, giving the collaboration a stated commercial destination as well as a technical assignment.

That relationship deserves attention from a factory buyer. An industrial automation proposal has to make sense within an existing operation, where equipment, processes and people already have responsibilities. A partner with industrial domain knowledge can help frame the task the system must perform and the environment in which it must perform it. The investment and partnership connect funding for development with an organization that also plans to bring the results into customer operations.

The agreement remains a plan for joint development and delivery. It supplies no basis for assigning revenue, adoption rates or customer savings to the Series F. Its significance is more concrete: Hitachi has publicly described how it expects to carry the resulting technology into its industrial offering. As Agile expands, the relationship gives both companies a shared piece of work that reaches from model development toward customer implementation. The commercial promise will develop through the applications they can put into service.

NVIDIA links the research workflow to the hardware portfolio

NVIDIA participates in the new round while already working with Agile on a product aimed at researchers. Agile ONE SN, announced October 21, combines a compact humanoid with NVIDIA Isaac GR00T software and Jetson Thor computing. The company schedules availability for early 2027, placing the product within the next phase of development rather than treating it as an installed fleet today.

Its proposed value is practical for a research team: bring demonstration capture, simulation, training, evaluation and deployment into a more connected workflow. Each handoff between those activities can require integration work before the team gets to study the robot skill itself. An integrated development system gives Agile another relationship with the people creating capabilities that may eventually reach industrial equipment.

Agile describes ONE SN as the research counterpart to its full-size industrial humanoid, Agile ONE. The product strategy makes a useful distinction between the environment in which a skill is developed and the environment in which a customer expects it to work. A research platform can make experimentation easier while the industrial application still requires its own evaluation. For buyers, that leaves room to ask how a demonstrated capability will transfer into the specific production task they need performed, with the surrounding workflow and people included.

The new capital carries manufacturing and customer assignments

The Series F also funds physical expansion. Agile says its Fürstenfeldbruck facility, opened early in 2026 near Munich, has annual capacity for up to 10,000 legged and wheeled humanoids. Capacity describes what the site is designed to produce; it should remain separate from delivery figures. The company's reported base of more than 20,000 installed robotic solutions spans its broader industrial portfolio.

That combination puts two different scaling problems inside the financing. More manufacturing capacity can support availability as demand develops, while new markets require relationships and applications that help customers put the equipment to work. The public announcement gives the direction of that spending without providing a production-utilization figure, a humanoid shipment count or a timetable for the US expansion.

SoftBank's involvement extends across that development history. A company-issued September 2021 release recorded a $220M Series C led by SoftBank Vision Fund 2. The new round now sits beside a broader set of research and industrial relationships. As Chen's team builds its presence in the US, the work includes finding where those relationships can make a factory deployment easier to develop, explain and support. Each successful application would give the company another setting in which machinery, customer knowledge and the next training cycle can meet.