Lyte, a Mountain View startup founded by the engineers behind Apple's Face ID depth-sensing technology, has raised $165 million in Series C funding at a $1.6 billion post-money valuation, according to Tech Funding News. The round, led by Maverick Silicon, more than triples the company's valuation from January 2026, when it emerged from stealth with $107 million already raised and a pitch to give robots a reliable way to see the world.
Total funding now stands at $272 million, and the company says it is moving from prototypes into production deployments with customers in inspection, logistics, and manufacturing. As coverage of breaking AI news has shown over the past year, the race to supply the physical economy with AI is increasingly about the unglamorous layers beneath the models — and perception is emerging as one of the most contested. For more context on this story, see our ongoing more AI stories.
A Reunion of the Face ID Depth Team
Lyte was founded in 2021 by Alexander Shpunt, Arman Hajati, and Yuval Gerson, three former Apple engineers who worked on the depth-sensing technology behind Face ID. Shpunt's pedigree in 3D sensing runs deeper still: he previously co-founded PrimeSense, the Israeli company whose technology powered Microsoft's Kinect before Apple acquired it in 2013.
The company emerged from stealth in January 2026 with LyteVision, its perception platform. Rather than leaving robot makers to stitch together separate camera, motion, and depth feeds, LyteVision fuses a camera, inertial motion sensing, and what the company describes as a 4D sensor that measures distance and velocity into a single synchronized data stream. The pitch is that by measuring motion natively in silicon and software, the fusion problems that cost warehouse robots time and accuracy never appear downstream.
Why Investors Are Betting on Silicon, Not Just Software
Maverick Silicon led the round, and its managing partner Andrew Homan has joined Lyte's board. The fund has previously backed edge-AI chip firm SiMa.ai and TensorWave. Existing investor Fidelity Management & Research, which led Lyte's Series B, returned alongside Atreides Management, Key1 Capital, and Ora Global.
"Physical AI has a sensing problem before it has a model problem," Shpunt said in a statement reported by Tech Funding News. "A robot cannot act safely on data that does not faithfully describe the world. We build the entire perception foundation, from custom silicon to the spatial data our systems produce."
Homan framed the bet in category-defining terms, saying Lyte is building "a foundational sensing platform that enables a wide range of robots to perceive and understand the world around them." Avigdor Willenz, the company's chairman, argued that owning the silicon is what separates companies that define a category from those that merely participate in one.
A Crowding Race for the Robot Eye
Lyte is not alone in chasing the perception layer of physical AI. According to the report, Rhoda AI raised $450 million at a $1.7 billion valuation in March, betting on video-trained foundation models rather than owning sensor silicon. RealSense, spun out of Intel with $50 million, is attacking the same camera-and-depth market from a different angle, while Luxonis raised a $14 million Series A in July around its open-source OAK camera line.
The differentiator, in Lyte's telling, is vertical integration: where rivals license models or sell modular hardware, Lyte owns the full stack from custom silicon up, which it argues is what actually eliminates the latency and calibration problems robotics teams complain about. Whether that bet beats the video-trained, license-everything approach will not be settled by a funding round, but by which approach ships fewer miscalibrated robots.
The Market Context
The funding lands amid rapid growth forecasts for AI-powered robotics. The AI robotics market is projected to reach $125 billion by 2030, according to figures cited in the report. Humanoid and warehouse robots remain expensive and error-prone in part because perception — reliably turning raw sensor data into a faithful model of the physical world — is unsolved at commodity price points.
Lyte says the new capital will go toward scaling silicon production, expanding LyteVision deployments, and growing headcount across silicon, software, optics, and manufacturing engineering.
What It Means for the Physical AI Stack
The round is a signal of where venture capital believes value will accrue in physical AI. The past two years produced enormous rounds for foundation models and for humanoid robot makers, but the 2026 funding cycle is increasingly rewarding the picks-and-shovels layers: simulation, sensor fusion, edge compute, and actuation.
For enterprises deploying robots in factories and warehouses, more capital in perception could translate into faster pilots converting to production — the step where many robotics programs have historically stalled. For the AI industry at large, Lyte's thesis is a reminder that embodied systems demand reliability standards that chat interfaces never confront: a robot acting on a misread scene does not produce a bad paragraph, it drops a pallet.
Lyte declined to name its production customers, according to the report. The company says deployments now span inspection, logistics, and manufacturing, and that the Series C will accelerate how quickly its perception stack reaches new robot types and environments.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
