A robot startup just raised $165 million to fix the one thing stopping machines from seeing the world clearly

Key points

• Lyte AI raised $165 million in Series C funding in September 2026, valuing the company at $1.6 billion.
• Maverick Silicon led the round, with Fidelity Management and Research Company, Atreides Management, Key1 Capital and Ora Global also participating.
• Lyte has raised $272 million in total since founding in 2021, including $107 million announced in January 2026.
• The company ships its LyteVision platform, a sensor system built on its own custom chips, to customers in logistics and manufacturing.
• LyteVision won a CES 2026 Best of Innovation award in Robotics.

Most robots today are assembled from parts that were never designed to work together. A camera from one supplier, a motion sensor from another, software trying to stitch it all into a coherent picture. The result, according to Sunnyvale-based Lyte AI, is a robot that can't fully trust what it sees.

Lyte's answer is to own the whole stack.

What does Lyte actually build?

Lyte designs processors, sensors and software so a robot understands its surroundings as one unified system rather than a patchwork. Think of it like the difference between a smartphone camera and a collection of loose parts held together with tape.

The centrepiece is LyteVision, a perception platform, meaning the full hardware-and-software stack that gives a robot its sense of the physical world. LyteVision combines 4D coherent vision (a technique that captures depth and motion simultaneously, similar to how radar sees through rain), high-resolution standard imaging, and inertial sensing (a chip that tracks acceleration and rotation, the same technology that tells your phone which way is up). All three run on Lyte's own custom silicon, a purpose-built processor chip, so the data from each sensor arrives on a single, perfectly synchronised timeline. The company says it owns the design from the transistor up.

"Physical AI has a sensing problem before it has a model problem," said Alexander Shpunt, Lyte's CEO and co-founder. "A robot cannot act safely on data that does not faithfully describe the world."

The practical payoff: a robot on a factory floor or in a busy warehouse can react faster and more reliably because it isn't waiting for software to reconcile disagreeing sensors.

Why does the funding matter?

The $165 million Series C, first reported by The Robot Report, gives Lyte the money to scale manufacturing and get LyteVision into more robots quickly. Maverick Silicon led the round; Fidelity Management and Research Company, which had led the Series B, returned alongside Atreides Management, Key1 Capital and Ora Global.

Lyte is already shipping to customers, so this isn't a research project waiting for a product. The cash goes to production capacity, expanded AI capabilities, and a hiring push across silicon, optics and software. Our 26 August story on Perceptron's warehouse-vision work showed how crowded the robot-perception space is getting; Lyte's custom-silicon angle is the sharpest differentiator we've seen.

Andrew Homan, managing partner at Maverick Silicon, cited the breadth of early customer demand as evidence that Lyte could become one of the defining companies of the robotics era. Homan has joined Lyte's board.

What does this mean for ordinary people?

Robots that perceive the world more accurately are safer to work alongside and more useful in warehouses and on factory floors. Better sensing is the unglamorous prerequisite for machines that can genuinely help rather than occasionally surprise the humans around them.

Lyte's progress is a useful reminder that the bottleneck in real-world robotics often isn't the AI brain. It's the eyes.

Common questions

Is Lyte selling a finished robot?

No. Lyte sells the sensing platform that goes inside robots built by other companies. Think of it as a supplier of eyes and ears, not the whole body.

Does this affect consumer products any time soon?

Not directly. Lyte's current customers are industrial: logistics firms, manufacturers and inspection companies. Consumer applications, if they come, are likely several years away.