While Jay Li advises caution against legal actions from major employers when launching a startup, he views the recent settlement with Tesla as a positive turning point for Proception.

“It feels like a resilience test, a pressure test,” Li told TechCrunch in an exclusive interview. “They say what doesn’t kill you makes you stronger, and that’s what I believe.”

Li, formerly the technical lead on Tesla’s Optimus humanoid robot initiative, faced accusations last year of improperly taking trade secrets to launch Proception. After months of legal back‑and‑forth, the parties reached a settlement that ended Tesla’s lawsuit earlier this month (Tesla declined to comment).

Li is now focused on solving what he sees as an even greater challenge: creating robotic hands that emulate human dexterity.

Proception announced Monday that it has closed an $11 million seed round led by First Round Capital, with investments from Y Combinator and BoxGroup.

The company also revealed that it is shipping the first batches of its high‑dexterity robotic hand to researchers and robotics firms, while opening the door to broader commercial orders. Li said the goal is to become the preferred hand supplier for companies that lack the resources or expertise to develop dexterous manipulation in‑house.

In the rapidly expanding robotics sector, Li feels that insufficient emphasis has been placed on replicating the intricacies of human hands.

Tesla CEO Elon Musk has highlighted the development of robotic hands as one of the most significant engineering hurdles remaining. While Musk believes Optimus could start working in factories within a few years, broader consensus suggests that achieving human‑level dexterity will take many more years. Kevin Lynch, director of Northwestern University’s Center for Robotics and Biosystems, estimated a decade before such systems become functional and useful enough to perform tasks humans routinely do.

Li argues that Proception can accelerate this timeline, largely due to its data collection methodology.

Most current humanoid training employs teleoperators: a human wearing a virtual‑reality headset observes the robot’s perspective and manipulates objects, with the robot learning from those commands.

Li points out that this approach lacks tactile feedback for the operator and is limited by the number of robots a company can run at one time.

Proception’s alternative is a sensor‑laden glove. Human testers wear the glove (and a headset), allowing Proception and its clients to capture “human hand interaction data without requiring a robot in the loop,” the company explains in its press release.

The same glove is mounted on Proception’s robotic hand, providing it with a sensor‑rich “skin.” The hand boasts 22 degrees of freedom and multiple joints per finger, enabling a wide range of dexterous movements.

Li says this technique also enables the company and its customers to obtain finer, task‑specific data, helping the robotic hand more accurately emulate human behavior. It also scales more efficiently.

“You need both hardware and data, and they must be developed together to achieve dexterous manipulation. Many companies focus only on hardware or on non‑scalable data collection,” Li said. “We’re combining highly sophisticated hardware with scalable data, which we see as the key to solving this problem.”

First Round partner Bill Trenchard, who led the investment, cited this approach as the main reason for backing Li.

“We believe they will have the best, possibly the most sophisticated hand on the market, backed by the data and models needed to support it,” Trenchard told TechCrunch. “Dexterous manipulation is essential for the future of humanoid robots, and it’s the final hurdle to achieving high performance.”

Trenchard added praise for Li’s composure during the lawsuit, describing him as a strong leader who kept his team focused.

Li expressed confidence that, after confronting Tesla’s formidable litigation arm, he would not be surprised if Tesla reaches out for collaboration as Proception expands.

“I think it will happen,” Li said.

Source link

Exit mobile version