Wat is er gebeurd
Daimon Robotics introduced Daimon‑TWM, a tactile‑grounded world model (TWM) designed to give robots a richer sense of physical interaction. The model combines tactile sensing hardware, multimodal data collection, and AI‑driven physical cognition to predict contact outcomes and adjust robot actions in real time. Demonstrations featured tasks such as handling glass pieces, preparing fruit skewers, and organizing items inside a refrigerator. The launch follows two financing rounds this year, including a strategic round led by Ant Group.
The article, published on September 29, 2026 by Robotics & Automation News, states that Daimon Robotics has launched Daimon‑TWM, a tactile‑grounded world model intended to enable robots to interpret physical interactions such as force, friction, deformation, and slippage. The company describes the model as a "tactile‑grounded world model" that uses touch as a core source of information for physical intelligence, combining physical cognition, predictive decision‑making, and real‑time control.
Daimon‑TWM builds on the firm’s earlier Vision‑Tactile‑Language‑Action (VTLA) architecture, which integrates tactile data alongside vision and other inputs. The new model aims to use tactile signals not only after contact but throughout the manipulation process, allowing robots to anticipate how an interaction will develop and adjust actions accordingly.
Demonstrations highlighted contact‑intensive tasks, including handling pieces of glass, preparing fruit skewers, and organizing items inside a refrigerator. The company attributes these capabilities to a combination of its vision‑based tactile sensors, a multimodal data‑acquisition initiative called Daimon‑Infinity, and the AI model itself.
The launch coincides with rapid financing activity: a strategic round led by Ant Group raised several hundred million yuan, following a 100‑million‑yuan Series A round two months earlier. Existing investors include China Merchants Capital, Lenovo Capital, Inovance Industrial Investment, China Mobile, and China Telecom.
Brongegevens: roboticsandautomationnews.com ↗
Waarom het ertoe doet
Robotic manipulation has long relied on vision, which can locate objects but cannot reliably gauge the forces needed for safe handling. By integrating touch as a core input, Daimon‑TWM addresses a key limitation in current robot systems, potentially reducing damage to fragile items and improving efficiency in manufacturing, logistics, and service settings. If the model proves effective, it could accelerate the deployment of robots in environments where delicate or deformable objects are common, expanding the economic scope of automation. The approach also highlights the growing importance of tactile data infrastructure, a relatively under‑developed area compared with large visual datasets, and may spur further investment in multimodal AI for embodied agents.
Current robotic systems often rely on visual perception to locate objects but lack reliable force estimation, leading to either excessive force that damages items or insufficient force that causes slippage. By grounding AI models in tactile data, Daimon‑TWM directly addresses this gap, potentially enabling more reliable dexterous manipulation across a broader range of objects and materials.
The model’s ability to predict contact outcomes and adjust actions in real time could reduce the need for extensive trial‑and‑error programming, lowering integration costs for manufacturers and expanding the range of tasks that can be automated. This is especially relevant for industries such as electronics assembly, food processing, and logistics, where handling delicate or irregular items is common.
The initiative also underscores the scarcity of large‑scale tactile datasets, a bottleneck for training physical AI models. Daimon’s effort to build a tactile data infrastructure through Daimon‑Infinity may encourage other firms to invest in similar data collection pipelines, accelerating progress in embodied AI research.
Interactief mechanisme: hoe het eigenlijk werkt
Ontdek interactief de onderliggende technologie achter deze ontwikkeling.
crm_get_transaction(id='4092').In AI, what are a model's "parameters"?
Wat je nu moet bekijken
Key questions include when and how Daimon‑TWM will be made available to customers, the pricing model, and whether the technology will be offered as a standalone AI service or bundled with Daimon’s vision‑based tactile sensors. Follow‑up reports should monitor performance benchmarks against existing force‑feedback systems, adoption by manufacturers, and any partnerships that could accelerate data collection for the tactile AI model. Additionally, watch for regulatory or safety standards that may emerge as tactile AI becomes more prevalent in industrial robotics.
Availability and pricing: The article does not disclose when Daimon‑TWM will be commercially available, nor the cost structure. Future announcements may clarify whether the model will be sold as a software license, bundled with hardware, or offered via a cloud service.
Performance validation: Independent benchmarks comparing Daimon‑TWM to existing force‑feedback or vision‑only systems will be essential to assess real‑world benefits. Look for third‑party evaluations or pilot deployments in manufacturing settings.
Partnerships and ecosystem development: Adoption may depend on integration with existing robot platforms and sensor manufacturers. Partnerships with major robot integrators or OEMs could accelerate market penetration.
Regulatory and safety considerations: As tactile AI enables robots to apply variable forces, safety standards may evolve. Monitoring any emerging guidelines from standards bodies will be important for commercial deployment.