The latest production setback for Elon Musk’s Tesla Optimus humanoid robot has nothing to do with AI bottlenecks or battery supply problems, but everything to do with hands. Not your average, run-of-the-mill hands, but hands that are capable of a high degree of freedom, also known as dexterous hands, which can imitate human manipulation abilities. This is called robotics’ “last mile problem,” which has proven to be a tough challenge to overcome, an IEEE Spectrum report revealed.
Dexterity needs the convergence of multiple technologies: touch sensors for the gripping function, kinesthetic sensors for spatial information, and actuation that is strong yet precise. As mentioned in industry studies, human dexterity requires the simultaneous control of multiple joints, precise torque delivery, and the ability to generalize in real time to novel objects. This has manifested at Tesla in not one but several redesigns of the forearms and hands of Optimus.
While Tesla is grappling with such and such complexities, a Chinese firm, InTime Robotics, has quietly established a strong lead in dexterous hands. Established by Cai Yingpeng in 2016, the essential product lines of InTime include micro servo electric cylinders and five-finger dexterous hands. The micro actuators developed by them make a precise replication of linear motion using a micron-level resolution of rotation, which facilitates high-force and compact actions. Unlike the first generation of linear actuators developed for toys, which had a plastic body and a low weight capacity, InTime’s innovations possess a load capacity of 10–20 kg load ratings which is a magnitude higher.
The journey of InTime is anything but linear. The firm changed gears from being an ODM robot maker to a component specialist because they noticed a gap: the insufficient supply of high-performance micro actuators in the sector. This indicated significant investment in the development of their own machining and research and development because it was hard to find companies that could provide the needed accuracy. The cost associated with the development of prototypes was above 10,000 RMB for each one of them.
The turning point came when InTime used the actuator technology to develop an all-metal dexterity hand. By incorporating micro servo electric cylinders in every joint of the fingers, the hand was made to possess a great deal of freedom to perform any task, whether it is assembling or handling. This is inline with the basics of advanced actuator engineering, which includes reducing backlashes, applying equal forces, and incorporating feedback control mechanisms in the system.
Commercial development remained low until 2019 when the company gained strategic funding from various financiers that saw the company through to the development of mass production maturity. Its initial major application came within the medical aesthetic sector where the company’s electric cylinders were used as precision rods within aesthetic equipment in a niche yet high-demand sector that proved the reliability of the concept.
Since 2020, InTime aggressively pushed into automation, 3C manufacturing, semiconductors, new energy, and scientific research. The adaptability of InTime’s actuation technology compact size, adjustable stroke lengths, and high IP protection could integrate into different systems. For robot technology, these actuation systems played an essential role in humanoid robots because size matters in this domain. InTime dexterous hands are now being delivered in quantities over 10,000 in a year, with the sales contributed by this module alone combing over half of the company’s overall sales in 2025.
The technologies behind these tools correspond with global research on the design of robotic end effectors. Advanced humanoid hands feature multimodal sensing capabilities, including force/torque fingertip sensors, skins, and proprioceptive encoders, for sophisticated manipulation capabilities. Actuation methods are also improving with the development of artificial muscle fibers for more natural actuation by others in the field. InTime’s design is also electromechanical but can be compatible with collaborative control systems driven by artificial intelligence. For Tesla, a supplier like InTime symbolizes a shortcut to a solution to the “invisible hand” problem. Musk’s emphasis, however, on Tesla’s internal development implies that a barrier to Optimus’s massive scale production lies in Tesla’s ability to either measure up to, if not far surpass, the expertise in sophisticated engineering demonstrated by InTime. For now, the bottleneck at least in humanoid robotics starts where the work begins with the fingertips.
