Gasgoo Munich- Daimon Robotics Robotics has recently closed a 100 million yuan Series A funding round, according to Gasgoo. The investment was jointly led by INOVANCE's industrial investment arm and China Telecom. Capital raised will be used to build ultra-large-scale datasets containing physical interaction information, accelerate the development of physical world models, and drive data flywheels and commercial closed-loops in real-world scenarios.
Incubated by a research team at the Hong Kong University of Science and Technology (HKUST), Daimon Robotics Robotics is built on a foundation of globally pioneering monochromatic light visual-tactile technology. The startup aims to establish a tactile infrastructure for embodied intelligence, constructing a comprehensive capability ecosystem centered on tactile perception and spanning the full chain of "hardware-data-model."
Image source: Daimon Robotics Robotics
Specifically, Daimon Robotics Robotics' core product lineup includes the world's first multi-dimensional, high-resolution, high-frequency visual-tactile sensor; a teleoperation data acquisition system featuring force and tactile feedback; a Vision-Tactile-Language-Action (VTLA) multimodal data acquisition software and hardware suite; and corresponding manipulation models.
On the data front, Daimon Robotics Robotics has established the world's largest distributed data collection network, sourcing directly from real-world operational scenarios. This network uses data containing physical interaction information from the real world as core "fuel" to power the training of physical world models.
In April, Daimon Robotics Robotics partnered with Google DeepMind and dozens of other leading global institutions to release Daimon Robotics-Infinity, the world's largest tactile, full-modal, physical-world embodied dataset. Within its first month of open-source release, the dataset garnered over 1 million downloads on Alibaba's ModelScope community, topping the charts for physical-world embodied datasets.
The Daimon Robotics-Infinity dataset relies on proprietary data acquisition equipment developed by Daimon Robotics, including two-finger grippers and five-finger gloves. These devices are equipped with visual-tactile sensors featuring 110,000 sensing units and a high frequency of 120Hz. Working in tandem with fisheye cameras, encoders, IMUs, and stereo cameras, the system provides comprehensive dimensional information for the dataset, covering tactile, visual, motion trajectories, executed actions, and voice text.
On the tactile front alone, it delivers high-density, full-modal information—including contact force, deformation, slippage, and object material properties such as topography, texture, and hardness. This data fills a critical gap in physical interaction features required for fine manipulation.
Image source: Daimon Robotics Robotics
To rapidly scale data volume and break through the spatial limitations of traditional, closed collection factories, Daimon Robotics has constructed a massive distributed embodied data acquisition network. Leveraging lightweight equipment and a distributed collection system, data collectors can access diverse, real-world application scenarios, enabling space-unrestricted data acquisition with an annual capacity reaching millions of hours. Through this network, Daimon Robotics-Infinity has established a global data co-construction system covering core scenarios such as industrial assembly, smart logistics, elderly care, home services, scientific research, catering and retail, and outdoor spaces.
According to the roadmap, Daimon Robotics-Infinity is set to expand its data scale to millions of hours within the year, comprising nearly 1 billion embodied data points. Of this, 10,000 hours of data will be open-sourced for the entire industry. Currently, the first batch of 1,000 hours of real-world data has gone live on Alibaba's ModelScope community.
Building on this foundation, Daimon Robotics recently partnered with Galaxy General to launch RobOmni, the industry's first tactile, full-modal evaluation benchmark for physical interaction capabilities that supports both "real data training" and "simulator training." Featuring a tactile full-modal simulation base, task evaluation protocols centered on contact-rich manipulation, and a complete Sim-to-Real verification chain, RobOmni aims to propel embodied intelligence from a "demo-driven" phase toward a "standard-driven" era.
About Seeds Discovery:
Gasgoo's "Seeds Discovery" column is designed to create a service platform connecting startups, industrial chain ecosystem partners, investment firms, and local governments, deeply empowering both the upstream and downstream sectors. Since its inception, the column has been dedicated to discovering high-quality companies, technologies, and business models that offer significant inspiration and leadership during the wave of intelligent transformation, thereby fostering the growth of innovative forces in the automotive industry. According to Gasgoo statistics, nearly all startups featured in "Seeds Discovery" have successfully connected with resources across the industrial ecosystem.
