Proception closes $11 million seed round after settling Tesla lawsuit

After resolving a trade secret lawsuit brought by the electric vehicle manufacturer last year, Proception, a robotics startup founded by former Tesla Optimus technical lead Jay Li, closed an $11 million seed round. First Round Capital led the round, with Y Combinator and BoxGroup also taking part.

Li founded Proception, a company that creates deft robotic hands for humanoid robots, after leaving Tesla, which accused him of stealing confidential information.

The lawsuit was dropped earlier this month following a settlement between the parties.

What Proception is building

Alongside the funding announcement, the company released ProHand 1.0. The robotic hand has a 22-degree-of-freedom tendon-driven design, which means that motors pull cables to move individual finger joints in a manner akin to how muscles and tendons function in a human hand.

According to Proception, hand surgeons provided input for the development of the device’s skin-like sensors, which detect contact during gripping and object handling. Proception is opening broader orders and shipping the first ProHand units to robotics firms and researchers this week.

According to the startup, computer vision and locomotion have advanced more quickly in the robotics sector than hands. Many robots still use simple grippers, which break down when tasks require precise pressure control or frequent repositioning of objects.

Li told TechCrunch that the key to solving dexterous manipulation is to combine hardware with scalable training data. The majority of businesses train humanoid robots via teleoperation, in which a person using a virtual reality headset controls a robot’s movements from a distance. Li sees the number of physical robots available as a bottleneck for data collection, and operators do not receive tactile feedback from the objects the robot touches.

ProGlove, a wearable with the same sensor skin as ProHand, is Proception’s substitute. To record interaction data without a robot in the loop, researchers wear the glove with a headset and perform tasks with their own hands. According to The AI Insider, robotic systems can then be trained on fine motor tasks using that data.

Li told TechCrunch, “You need both hardware and data, and those need to come hand-in-hand. We’re working on this highly dexterous hardware plus highly scalable data. We believe that’s a key combination to solve this problem.”

The difficulties Proception encounters are widely known. The Wall Street Journal was informed last year by Kevin Lynch, director of Northwestern University’s Center for Robotics and Biosystems, that functional robotic hands that are comparable to human hands are probably ten years away.

Elon Musk, the CEO of Tesla, has stated that one of the most significant unresolved engineering issues in the industry is robot hands.

If you're reading this, you’re already ahead. Stay there with our newsletter.