Tesla Settles Trade Secret Fight With Ex-Engineer’s Robot Hand Startup That Just Raised $11 Million

Jay Li doesn’t recommend getting sued by Tesla. The experience tests you. It pressures every part of a young company. Yet here he sits. Proception, the robotic hand maker he founded, has settled the high-profile dispute. Tesla dismissed its lawsuit earlier this month. And on the same day the news broke, the startup announced an $11 million seed round. Investors lined up anyway.

Li led technical work on Tesla’s Optimus humanoid program before he left in September 2024. Bloomberg first reported that Tesla accused him of downloading sensitive files about advanced robotic hand sensors in the weeks before his exit. The company claimed he used that information to launch Proception just days later. The designs looked too similar. The timing raised red flags. But months of legal fighting produced no injunction. A federal judge in San Francisco turned down Tesla’s request last fall, saying the carmaker failed to show strong enough odds of winning on the merits even though it demonstrated likely harm. The case dragged on until both sides reached terms. No details of the settlement emerged. Tesla offered no comment.

“People say that what doesn’t kill you makes you stronger, right?” Li told TechCrunch in an interview published Monday. He described the ordeal as a resilience test. A pressure test. The kind that either breaks a founder or hardens resolve. For Li it seems to have done the latter. Proception now ships its first high-dexterity robotic hands to researchers and robotics companies. Wider orders just opened. The fresh capital will accelerate that push.

The $11 million seed round came from First Round Capital leading with participation by Y Combinator and BoxGroup. Bill Trenchard, partner at First Round, made clear why his firm backed the company even after the lawsuit cloud. “We think they will have the best hand in the market,” he said. Dexterous manipulation stands as one of the final barriers. The last mile before humanoid robots deliver real performance in factories, homes, or warehouses. Many companies would rather buy proven hands than build their own from scratch. Proception aims to become that supplier.

Hands remain brutally difficult. Elon Musk has called them among the toughest engineering problems in the entire Optimus project. Kevin Lynch, a professor at Northwestern University, told reporters it could take another decade before robotic hands become truly functional and useful in everyday settings. The hardware must replicate dozens of degrees of freedom. Fingers need multiple joints. Skin-like sensors must detect subtle forces. Control systems have to coordinate everything at high speed. Training data for the AI that drives these motions proves even harder to gather at scale.

Proception attacks the data problem with a novel twist. Its system uses a sensor-packed glove and headset. Human testers wear the gear and interact with objects in the real world. No robot sits in the loop during collection. The approach captures finer, task-specific movements. It generates volumes of high-quality information that can train manipulation models more effectively. Li believes this combination of capable hardware and scalable data gives his company an edge. “You need both,” he explained.

The hardware itself boasts 22 degrees of freedom. Multiple joints per finger allow a wide range of dexterous motions. That specification puts it in rare company among commercial offerings. Most grippers on the market today remain simple two-finger clamps or basic three-finger designs. They handle pick-and-place but struggle with anything requiring human-like touch or adaptability. Proception’s hand targets the gap. Early customers include academic labs and humanoid startups that want to focus their own engineering resources elsewhere.

And the timing feels pointed. Humanoid robots have moved from science fiction to factory pilots in just the past two years. Figure, Boston Dynamics, Agility Robotics, and Tesla itself all chase the same prize. Each needs reliable hands. Few have solved them. Tesla’s own Optimus prototypes still rely on hands that Musk has repeatedly described as the biggest remaining challenge. Some observers on X noted Monday that the settlement removes a lingering legal shadow for Proception while Tesla’s stock climbed more than 7 percent on broader market strength. The lawsuit itself did not appear to drive the move. Yet it underscored how fiercely the carmaker guards its humanoid technology.

Li sounds unfazed by the past conflict. He even suggested Tesla might one day become a customer. “I wouldn’t be surprised if it happens,” he said. “I think it will happen.” Confidence like that can sound bold after a trade secret battle. But investors clearly bought the vision. First Round’s involvement signals serious belief in the technical approach. YC’s participation adds credibility given its deep robotics network. BoxGroup brings early-stage discipline.

The broader robotics investment climate has heated up again. Just weeks ago TechCrunch reported that Barcelona-based Theker closed an $85 million Series A to build adaptable factory robots. Capital flows toward any team that can demonstrate progress on manipulation, locomotion, or low-cost production. Proception’s ability to ship hardware now, rather than promise prototypes later, gives it tangible proof. Researchers can start testing the 22-degree-of-freedom hand immediately. Data from those deployments will feed back into Proception’s own models. The flywheel begins to spin.

Trade secret suits in Silicon Valley often end in quiet settlements. This one carried extra weight because it touched the Optimus program, one of Tesla’s most secretive efforts. The complaint had detailed how Li allegedly copied files onto personal devices. It pointed to the striking resemblance between Proception’s early demos and Tesla’s internal designs. Yet the court found Tesla’s evidence insufficient for an injunction. That ruling likely shaped the eventual compromise. Both sides avoided a trial that could have exposed more technical details in open court. Now Proception moves forward unencumbered. Tesla keeps its remaining secrets.

Challenges remain. Manufacturing hands at scale with consistent quality will test Proception’s operations team. Competition could intensify if larger players decide to open their own supply chains. And the AI models that turn raw sensor data into useful behaviors still require enormous compute and clever algorithms. Li’s glove-based data collection may help, but it must prove superior in real deployments. Early feedback from the first shipped units will matter greatly.

So the story isn’t over. It has simply shifted from courtroom to laboratory. From legal filings to product roadmaps. Proception now holds working hardware, fresh capital, and a founder who survived one of the industry’s most aggressive enforcers. The robotics community will watch closely. If these hands perform as promised, they could accelerate multiple humanoid programs at once. Companies that once feared building their own dexterous grippers might instead buy from a former Tesla engineer. The ironies run deep. But in this corner of technology, results speak louder than past disputes.

Li’s closing thought carries weight. The lawsuit tested everyone involved. It slowed momentum at a critical moment. It scared some potential limited partners. Yet the team kept building. The round closed. The hands shipped. That combination of legal survival and product delivery stands out in a field crowded with vaporware. Investors bet that Proception can turn its hard-won resilience into market leadership. The coming months will test whether that bet pays off.