Daimon Robotics building the infrastructure layer for tactile intelligence

Credit: Daimon Robotics Vision has become central to modern robotics, but cameras alone cannot tell a robot everything it needs to know about physical contact. When handling flexible objects, inserting components, or manipulating delicate items, robots need information about forces, slippage, and contact that visual systems may miss.

Daimon Robotics wants to provide the infrastructure to help robots develop that capability. Rather than focusing exclusively on tactile sensors or developing a single foundation model, the company is building a pipeline that connects tactile sensing, data collection, simulation, and AI models.

Its goal is to turn physical interactions into useful training data that helps robots perform manipulation tasks more reliably. “By building tactile intelligence, we’re actually playing the role as an infrastructure provider,” Daimon’s Partner & VP of Technology Dr. Zhu told The Robot Report .

“We’re not aiming to be a foundation model developer. We are providing the infrastructure for doing so for those downstream customers.” That distinction is central to Daimon’s strategy: develop the tools and data infrastructure that other companies can use to build more capable robotic systems.

Building a pipeline for tactile data Collecting useful tactile data at scale presents a challenge. Unlike visual data, which can be captured relatively easily with cameras, tactile data depends on physical contact. The resulting information varies with an object’s material, geometry, the forces applied, and the task being performed.

Daimon is developing a pipeline to capture and process these interactions, with tactile feedback operating at 30 Hz. The company is also working to connect real-world data with simulation, allowing physical interactions to inform training in virtual environments. The sim-to-real gap remains a significant challenge, particularly when accurately reproducing physical properties such as friction and contact forces.

Daimon’s approach starts by calibrating simulations against real-world measurements before transferring collected data into simulated environments. “Let’s be frank; it’s not good enough. It’s kind of terrible,” Zhu said of the current state of simulation, referring to the gap between simulated and real-world environments.

Daimon is developing tactile simulations that account for physical properties measured through real-world testing. The goal is to improve the fidelity of those simulations and make it easier to transfer learned capabilities to physical robots. The company has discussed collecting more than 10,000 hours of data, with its longer-term Daimon-Infinity initiative targeting more than 1 million hours of tactile embodied data.

Zhu also emphasized the importance of working with companies developing robotic models and applications. “Doing everything ourselves was not enough,” he said. “Collaborating internationally in different downstream scenarios is really important.” That collaboration helps Daimon understand what other developers need from its sensors, data-collection devices, software, and evaluation tools.

Credit: Daimon Robotics From tactile demonstrations to industrial applications Daimon is targeting manipulation tasks such as assembly and packaging, where robots must respond to physical contact rather than simply follow predetermined trajectories. A robot might need to determine whether a component is properly seated, detect when an object slips, or adjust its movements when a part behaves unexpectedly.

Tactile feedback can help it respond to these conditions. The company demonstrated applications involving delicate manipulation at IROS 2026, including applying heat-transfer designs to tote bags and threading bracelets. It also showed a robot collecting and packaging broken glass, illustrating how models trained on data from varied scenarios can be prompted to perform specific tasks.

Daimon uses its Umi device to collect data from human interactions. Zhu said the company’s demonstrations were powered by foundation models trained on data collected across multiple scenarios, rather than being built solely around individual robot demonstrations. The company has identified Lenovo Capital and Inovance Technology as key partners.

Zhu said Daimon has also moved beyond proof-of-concept projects into industrial assembly lines and warehouses, although he declined to name specific customers because of nondisclosure agreements. The commercial test is whether tactile sensing and the models built around it can improve performance consistently in production environments, where robots must handle variations in objects, materials, and operating conditions.

Moving beyond visual perception Daimon’s broader strategy centers on developing models that can use tactile information to understand and respond to physical interactions. Its Daimon-TWM effort focuses on a tactile-grounded world model intended to help robots reason about physical interactions, rather than relying on visual perception alone.

Recognizing an object is only part of manipulation. A robot must also understand how its actions affect that object and adjust when the physical result differs from expectations. Tactile feedback can help close that loop by providing information about contact and the consequences of an action.

Zhu described two roles for tactile sensing in robotic systems: helping models reason about physical properties and enabling rapid adjustments during contact-intensive tasks. The latter can require high-frequency feedback, with Daimon targeting control at 30 Hz. The company is also exploring how data collected through human interactions can help train models for different robot configurations, including humanoids and dual-arm manipulators.

But collecting data and training models are only part of the problem. Developers must still demonstrate that these systems generalize across different objects, materials, robots, and operating conditions. Scaling touch for robotics The broader question facing the robotics industry is what infrastructure is needed to make physical AI work reliably outside controlled demonstrations.

Companies are investing heavily in vision-language-action models, simulation, and large-scale robot training datasets. Touch adds another dimension, particularly for robots that must manipulate objects through physical contact. Daimon aims to build part of the infrastructure connecting sensing, data, simulation, and intelligence.

Its strategy is to support the companies developing robotic systems rather than compete directly with them to build the next general-purpose robotic brain. Its success will depend on whether that approach measurably improves manipulation, reduces the effort required to develop new applications, and supports deployments across different industrial environments.

Zhu was careful not to overstate the capabilities of today’s models. Even when trained on large datasets, he said, their ability to generalize remains limited. “The model is not as intelligent as we think,” Zhu said.

“It has some generalization ability, but it was in-domain generalization sometimes.” He added that embodied-intelligence models are still in development, and that collecting data from similar tasks can help them generalize to new applications. For Daimon, the challenge is turning that emerging capability into reliable performance across a broader range of real-world tasks.