New Dexterous Hands for Robots: Compute Services Power the Next Leap

Introduction: The Rise of Specialized Robotic Hands

As embodied intelligence advances in motion control, environmental perception, and human-robot interaction, the industry now faces a more concrete question: can robots truly perform real work in factories, labs, and homes? On August 7, 2025, Zhongke Huisi (Hunan) Co., Ltd. (Zhongke Huisi) made its debut in Changsha, unveiling three dexterous hand products—L1, D1, and M1—along with a pre-released F-series tendon-driven bionic hand. The company also introduced data gloves and a data collection toolchain. This launch marks a strategic move to build a comprehensive dexterous manipulation capability system, underpinned by advanced compute services and edge AI.

The Product Lineup: Diverse Designs for Different Tasks

Zhongke Huisi's multi-series approach targets distinct operational requirements through different technological routes. The L1 model emphasizes lightweight construction and simplified degrees of freedom (DoF), prioritizing cost-effectiveness, stability, longevity, and ease of deployment. Its streamlined DoF configuration reduces control system complexity and cost, while the lightweight design minimizes the load on robotic arms. Fast open-close response enables high-frequency actions and continuous operation. At the WAIC event, a robot band named CASBOT BAND, equipped with L1 hands, performed live with four robots playing guitar, bass, keyboard, and drums.

The D1 is a high-DoF universal five-finger hand, focusing on movement flexibility, independent multi-joint motion, multi-finger coordination, and tactile sensing. The M1 is a modular scenario-specific hand, designed for complex operational tasks, featuring a 'standardized body plus functional fingertips' architecture for rapid switching between different capabilities.

The pre-released F-series explores high-DoF, biomimetic structures, and compliant manipulation through tendon-driven mechanisms, aiming to replicate human-like hand movements with high flexibility and soft interaction.

Beyond Specs: A Focus on Real-World Performance

Zhongke Huisi deliberately avoids the race for higher specifications. Instead, it aims to establish a complete logic from hardware to data to skills. The core outcome of this launch is not a single product but a comprehensive dexterous manipulation capability system for the embodied intelligence era. The company proposes a 'dual-flywheel' development model: on one side, a series of dexterous hand bodies with different structures, DoF, and application positioning provide hardware foundations for robots, research, data collection, and applications. On the other side, data gloves, teleoperation devices, and data collection toolchains continuously accumulate dexterous manipulation data, driving iterative improvements in models, skills, and hardware.

Market Dynamics: Rapid Growth and Investment

According to the '2026 H1 China Dexterous Hand Industry Research Report' by GGII, China's dexterous hand market sold approximately 19,200 units in 2025, a year-on-year increase of 236.84%. Projections indicate sales will reach 70,200 units in 2026, up 265.63%, and exceed 430,000 units by 2030, with a compound annual growth rate of 57.79% from 2026 to 2030. Between 2023 and May 2026, the domestic industry saw about 57 financing deals totaling around 12.8 billion yuan. This surge in market demand and capital investment is driving dexterous hands from technical validation and small-batch use toward productization, scale, and real-world deployment. However, challenges remain in cost, performance, stability, and operational data accumulation.

The Role of Compute Services in Dexterous Manipulation

While the industry often evaluates dexterous hands based on parameters like DoF, weight, payload, speed, precision, and sensing configuration, these metrics alone do not guarantee operational skill. For instance, a robot can pick up a screwdriver, but performing a screwing action requires precise control of tool orientation, position, trajectory, and force. This is where compute services become critical. Real-time processing of sensor data, high-frequency control loops, and AI inference for adaptive manipulation demand robust computational resources. Edge computing platforms can provide low-latency processing at the robot, while cloud-based services can handle large-scale model training and data analytics. Zhongke Huisi's dual-flywheel model relies heavily on data collection and model training, which are compute-intensive tasks. The company's focus on building a 'professional platform' for dexterous manipulation inherently depends on advanced compute services to process the vast amounts of teleoperation data and refine AI models.

From Component to Platform: The Evolution of Zhongke Huisi

Zhongke Huisi was established in July 2026, emerging from the industry transformation. The company focuses on dexterous hands and embodied manipulation capabilities, aiming to evolve from a single hardware component to a software-hardware integrated operation capability platform. Its core philosophy is 'from a dexterous hand to a set of embodied manipulation capabilities.' Zhongke Huisi is not starting from scratch; it leverages the R&D capabilities and scenario practices of Beijing Zhongke Huiling Robot Technology Co., Ltd. (Zhongke Huiling), founded in 2023 by Zhang Zhengtao, who also serves as a researcher at the Institute of Automation, Chinese Academy of Sciences. The new company is jointly established by Zhongke Huiling, Lens Technology, and Hunan Huaxia Investment Group, combining expertise in robotics, precision manufacturing, and application scenarios.

As the team advanced, they realized that dexterous hands are no longer just an end-effector but a specialized field requiring independent development. The market potential is vast, and the operational data generated by high-DoF hands is invaluable for training robot 'brains' and foundation models. Zhongke Huisi has settled in Hunan Xiangjiang New Area and announced participation in building a specialized dexterous hand skills training ground, focusing on data collection, skill training, and application validation.

Conclusion: The Future of Robotic Manipulation

Zhongke Huisi's launch signifies a shift from parameter competition to capability building. By integrating hardware innovation with a robust data collection and training pipeline, the company is positioning itself as a key player in the embodied intelligence ecosystem. Compute services, particularly edge and cloud computing, will be the backbone of this transformation, enabling real-time control, advanced AI, and scalable data processing. As the market grows and technology matures, the collaboration between hardware developers and compute service providers will be crucial to unlocking the full potential of dexterous robotic hands in real-world applications.