A former Tesla engineer just turned a trade secret lawsuit into a launchpad. Jay Li, a onetime technical lead on Tesla’s Optimus humanoid robot program, has settled the suit his old employer filed against him and pulled in $11 million to build robot hands, according to TechCrunch AI. Tesla dismissed the case earlier this month and did not respond to TechCrunch’s request for comment.
Li’s company, Proception, accused last year of walking off with trade secrets, came out the other side intact. “I think it’s kind of like a resilience test, or pressure test,” Li told TechCrunch in an exclusive interview. “People say that what doesn’t kill you makes you stronger, right?”
What happened
Proception announced two things on Monday:
- An $11 million seed round led by First Round Capital, with money from Y Combinator and early-stage fund BoxGroup.
- The first shipments of its “high-dexterity robotic hand” to researchers and robotics companies, plus wider orders opening up.
Li’s goal is to become the top hand supplier for companies that don’t want to build dexterous manipulation themselves. That’s the industry term for getting a robot to handle objects the way human fingers do.
Why this matters
Money has flooded into robotics, but Li argues not enough of it has gone toward the hands. His old boss agrees. Elon Musk has called robot hands one of the biggest unsolved engineering problems, even as he promises Optimus could be working in factories within a few years.
The expert consensus is far more cautious. Kevin Lynch, director of Northwestern University’s Center for Robotics and Biosystems, told the Wall Street Journal last year that functional, genuinely useful robot hands are likely a decade away. So Proception is chasing what many call the last mile of the humanoid story, the piece that decides whether these machines are actually useful or just impressive demos.
The technical bet
Here’s where Proception’s approach stands out. Most companies training humanoid robots rely on teleoperators: a human in a VR headset sees what the robot sees and moves objects for it, and the robot learns from those commands.
Li sees two problems with that. The teleoperator gets no physical feedback from the objects being touched, and the whole method is capped by how many robots a company has on hand at any moment.
Proception’s answer is a sensor-packed glove. Human testers wear it along with a headset, capturing detailed hand-interaction data without needing a robot in the loop. The same glove then doubles as the “skin” on the hand Proception is building, a design with 22 degrees of freedom and multiple joints per finger.
“You need both hardware and data, and those need to come hand-in-hand to get [dexterous manipulation] to work,” Li told TechCrunch. “A lot of companies solely focus on hardware, or like hardware plus non-scalable data. We’re working on this highly dexterous hardware plus highly scalable data.”
What stands out to me is the data angle. Plenty of startups can machine a fancy hand. Far fewer have a cheap, scalable way to teach it. Decoupling data collection from physical robots is how Li thinks he beats that decade-long timeline.
The investor view
First Round partner Bill Trenchard, who led the round, isn’t hedging. “We think they will have the best hand in the market, maybe the most sophisticated hand today, and the underlying data and models to support that,” he told TechCrunch. He called dexterous manipulation “the last mile of getting these robots to be truly performant” and praised Li for keeping calm under Tesla’s legal pressure.
What to watch
With hands now shipping and orders open, the test is whether other robotics companies actually buy Proception’s hardware-plus-data pitch instead of rolling their own. If they do, Li becomes a picks-and-shovels supplier in one of the hottest corners of robotics.
Li is confident enough to predict his former employer eventually comes knocking. After facing down Tesla’s litigation team, he told TechCrunch he wouldn’t be surprised if the company asks Proception for help as it grows. “I think it will happen,” he said. Full details are at the original TechCrunch AI report.