TetherIA.ai发布新一代AI驱动智能机械手，革新硅谷创业生态

近日，硅谷初创公司TetherIA.ai发布了一款基于人工智能的下一代灵巧机械手技术，该技术通过先进的算法和机器学习技术，实现了机械手的精准控制和高效操作，这一创新将极大地推动机器人技术在医疗、制造等领域的应用和发展。

TMTPOST -- TetherIA.ai, a Silicon Valley-based robotics startup founded in early 2025, is making waves in the robotics industry with its development of a next-generation humanoid dexterous hand platform. Co-founded by veterans from Tesla and Waymo, the company seeks to push the boundaries of AI-integrated robotics for commercial applications.

The startup’s founders bring heavyweight experience from the tech world. Yiwei Tao, also known as Evan, led the Tesla Optimus Dexterous Hand project, overseeing everything from engineering design to mass production, and previously worked at Apple as the lead mechanical designer for Vision Pro. Co-founder Xu Dong, or Joe, formerly headed AI planning technology at XPeng and was a core member of Waymo’s foundational large model team, publishing several high-profile AI and autonomous driving papers.

TetherIA’s team currently includes 15 engineers, all based in Silicon Valley and drawn from top institutions such as Stanford, UC Berkeley, Cornell, and Carnegie Mellon. The group has prior experience at Tesla, Apple, Waymo, and XPeng, and specializes in robotics control, mechanical design, and AI.

The company’s first product—a five-fingered, tendon-driven humanoid dexterous hand—is already through initial R&D and hardware-software integration, though specifications remain under wraps. The hand is designed to be lightweight, modular, and highly durable, with human-like dimensions and a high degree of freedom. Its software includes an AI Copilot suite and a Sim-2-Real toolchain for teleoperation and other applications, aiming to bring practical embodied intelligence to a variety of tasks.

In a move to accelerate innovation within the robotics community, TetherIA is also launching an open-source dexterous hand priced around $300, positioned as one of the most advanced open-source platforms in the field. Both the commercial high-DOF hand and the open-source version will debut at the Conference on Robot Learning (CoRL) from September 27-30 in Seoul, South Korea, where the company plans to showcase real-time AI control for grasping and manipulating objects of varying sizes and shapes. Orders for the open-source hand will open immediately after the event, while the high-DOF model is slated for delivery early next year.

Image: Demonstration of tool usage capability

Image: Demonstration of fine object grasping capability

Image: Capability for continuous complex tasks

With global demand for flexible automation soaring, dexterous hands have become essential components for logistics, collaborative robots, precision assembly, and household service robots. According to Gaogong Robotics Industry Research Institute, global humanoid robot sales could reach 5 million units by 2035, with the market size topping $400 billion. Dexterous hands are estimated to account for 25-30% of the total bill of materials for these machines.

Investors see significant potential. Tong Ti, Investment Director at Matrix Partners China, said: “Dexterous hands are a key Tier 1 component for embodied intelligence and have a significant impact on the pace of its real-world adoption. The TetherIA team has the experience and capability to drive both hardware and software innovation, and we are pleased to support them as they bring their first product to market.”