A new robotics startup is setting the stage for the future of robotics, not by improving motors or batteries, but by fixing how machines see. Lyte has emerged from stealth with $107 million in funding and a sharp focus on one of the industry’s most stubborn constraints, including reliable, real-world perception.
Founders shaped by mass-market vision
Lyte was founded in 2021 by Alexander Shpunt, Arman Hajati, and Yuval Gerson, all former Apple engineers deeply involved in Face ID’s depth-sensing and perception systems. Shpunt’s influence runs even deeper. Before Apple, he co-founded PrimeSense, whose technology became the foundation for Face ID after Apple acquired the company in 2013.
PrimeSense also powered the Microsoft Kinect, one of the first computer-vision products deployed at massive consumer scale. This experience of shipping vision hardware that worked reliably outside the lab defines Lyte’s worldview.
The founders are building systems designed to operate in messy, unpredictable environments.
Building a visual brain for robots
Rather than offering another sensor, Lyte is positioning perception as infrastructure. Its flagship platform, LyteVision, combines three sensing modes: a camera, inertial motion sensing, and a 4D sensor that captures distance and velocity. The data is fused into a single system that produces immediately usable information, rather than raw signals that require extensive downstream processing.
Shpunt describes this as creating a “visual brain” for robots that handles both sensing and interpretation. The implication is strategic. If robots can see motion, depth, and context clearly in real time, safety decisions can be made faster and closer to the machine, reducing risk in dynamic environments.
Solving integration before scaling robotics
Lyte is also attacking a less visible drag on robotics adoption: integration friction. Robotics companies often spend years stitching together sensors, optics, silicon, and software from different vendors. Lyte’s answer is to build custom silicon, optics, and software into a single, plug-and-play stack.
This matters as the market accelerates. The AI robotics sector is projected to reach $125 billion by 2030, yet McKinsey & Co. reports that 60% of industrial companies lack the internal capability to implement robotic automation.
