Lyte Is Live

Today, we're introducing Lyte.

I'm proud of what this team has built and grateful to the people who made it possible. Over the past four years in stealth, our biggest accomplishment hasn't been a single milestone; it's been assembling an exceptional team and doing the hard work required to build a real perception platform for Physical AI.

In that time, we've raised over $107M, developed our core product, earned design wins with leading Physical AI companies, and were recognized at CES 2026 with Best of Innovation in Robotics and as an Honoree in Vehicle Tech & Advanced Mobility. Those outcomes matter, but they're a byproduct of the work, not the goal.

Lyte started in 2021, when our founding team—coming out of Apple—kept seeing the same pattern across robotics companies. Teams were spending years building perception stacks from scratch: sourcing sensors from multiple vendors, calibrating systems, debugging synchronization issues, and repeatedly solving problems we'd already encountered in consumer electronics at scale.

The sensing technologies themselves existed. What was missing was infrastructure.

At the same time, several things were changing. Silicon photonics and coherent sensing matured to the point where they could move from labs into production at robotics economics. Investment in Physical AI reached a level where serious systems were being built. Compute was no longer the limiting factor. Applications were ready.

Perception was still the bottleneck.

That gap is why we started Lyte: to build the perception infrastructure the industry didn't have.

Robotics companies shouldn't need to become experts in sensor integration, calibration, and synchronization. They should be able to focus on building robots, not reinventing vision systems.

Our team has been through this before: shipping perception technology at billion-unit scale, navigating manufacturing complexity, and making advanced sensing work reliably in real-world conditions. We're applying those lessons to Physical AI.

Lyte delivers complete perception platforms with plug-and-play integration—systems that operate reliably in fog, darkness, and bright sunlight, and that work alongside other sensors without interference.

When perception becomes infrastructure—when it simply works—the constraints change. Automation scales. Systems deploy faster. New classes of robots become viable. Progress stops being gated by how well machines can see.

Physical AI has reached an inflection point. Perception no longer needs to be the limiting factor.

If you're building robotics systems that need robust perception infrastructure, we should talk.

If you want to help build this future with us, we're hiring.

— Alex Shpunt
CEO & Co-Founder, Lyte