Switzerland has ranked first in the Global Innovation Index for 15 consecutive years. It is an important source of innovation in the world and China’s first innovation strategic partnership country. It is extremely complementary to China in the field of innovation and development.
The “Top 100 Swiss Innovations” hosted by Venturelab brings together many outstanding technological innovation companies.
The Mimic Robotics to be introduced in this article is a company listed in the “Top 100 Swiss Innovations” in 2025.
The company was founded in 2024 as a spin-out of ETH Zurich and was co-founded by Stefan Weirich, Elvis Nava, Stephan-Daniel Gravert and others.
Stefan Weirich is the company’s CEO and has extensive study and work experience in the field of mechanical engineering; Elvis Nava is the chief technology officer and holds a PhD in robotics from ETH Zurich; Stephan-Daniel Gravert holds a master’s degree in electrical engineering and information technology from the Technical University of Munich.
Although industrial automation has made great progress in the past few decades, conventional automated robots face many challenges in unstructured environments.
They are good at performing repetitive tasks in a highly controlled environment. Once faced with new tasks, they require expensive deployment and customized programming. They are prone to failure when the environment changes, and their flexibility and scalability are limited.
Humanoid robots also have limited application scope due to high cost and safety hazards.
However, it is predicted that by 2035, the global market size of humanoid robots and dexterous robots is expected to reach US$38 billion; by 2040, the broader robot market size is expected to reach US$200 billion to US$1 trillion.
Mimic Robotics has developed a universal humanoid manipulator based on physical AI models.
The robot is based on the company’s physical AI model and combines an AI-driven dexterous robot with a mature existing robot arm. It can perform various industrial tasks at a skill level close to humans, and has the advantages of high cost-effectiveness and rapid deployment.
Its core general physical AI model can be trained by collecting real-world factory data, solving the problem of lack of high-quality motion data in robot research.
The model is a scalable remote operating system that supports both glove-based and Apple Vision Pro control interfaces, and can quickly collect data for a variety of real-world tasks.
When used, multi-mode pre-training is first used to promote the robot’s learning tasks, and then the operator wears the data collection device to collect hand and arm movement data for training the AI model, so that the robot can accurately reproduce human operating skills.
Moreover, the model includes a robust data collection protocol to ensure that the robot autonomously responds to environmental changes, handles interference, and corrects its own behavior. The same algorithm is suitable for any application scenario.
The universal humanoid manipulator developed on this basis has 16 degrees of freedom, 21 joints, a 7kg payload and a fingertip accuracy of ±0.5mm.
It features a wide-angle wrist camera that provides multi-angle and low-latency visual feedback, supporting fine tendon-driven control.
Its human-like design has high compatibility with human data, can cope with harsh working environments, and accurately operate fragile objects. It can be equipped with one or two arms and installed on a flexible fixed workbench or an autonomous mobile platform.
Actual tests show that this robot can help enterprises reduce operating costs by 70% and can be deployed within one hour.
Mimic’s models and products are currently being piloted with top manufacturers, including Fortune 500 companies and global automotive brands, and have also established partnerships with leading multinational logistics providers.
In November 2025, Mimic completed a US$16 million seed round of financing. The funds will be used to accelerate the company’s development of universal physical AI models and universal humanoid manipulators, and promote cooperation and deployment with leading global industry companies.
