Luxonis raises $14M to develop the vision layer for intelligent automation

Denver-based Luxonis Holding Corp. announced Thursday it raised $14 million in early-stage funding to transform industrial automation with cameras and machine vision, providing a perception layer for robotics and automated systems to understand the real world.

Denali Growth Partners led the Series A round alongside participation from Denali Growth Partners.

Founded in 2019, Luxonis’ trajectory to this point is worth noting. The company started off this marathon on an uncommon path for tech startups: a 2020 Kickstarter campaign provided a $1.3 million foundation from over 6,500 backers. This has culminated today in cumulative capital raised of over $23 million, including the crowdfunding, according to Pitchbook.

Luxonis builds both the cameras and the software that give robots “vision,” a critical layer of Physical AI that enables AI vision-language models and vision-action models to perceive the real world. AI-driven robotics ingest data from numerous sources, including visual, to control actuators, robotic arms and other devices. Having high-quality inputs that can drive depth perception, multiple angles and signal processing can be the difference between being half-blind and having eyes fully open.

The company’s OAK Camera devices combine multiple vision sensors with on-device compute in a single unit, and its DepthAI open-source software ecosystem now enables developers to build agentic AI visual perception automation.

The latest generation of OAK4 launched in 2025, and the company said it will be scaling that rollout with this funding round. It comes in a variety of form factors for plug-and-play, providing superior vision and compute at the edge.

Luxonis said its cameras deliver high-accuracy depth with up to 1/32 subpixel precision and can run local AI models optimized for edge devices, compressed to INT8/FP16. This allows powerful acuity for robotics on factory floors and retail environments with high throughput and low power consumption, reducing the need for applications to reach out to the cloud for processing. It also means machines react faster and can make real-time decisions more quickly without the latency of running over the network and can continue to operate correctly even when the network is unavailable.

Right now, physical AI is very hot. AI models that run robots as “brains” and robotics hardware are being built to meet a swelling tide of interest in intelligent automation. Massive funding rounds have poured into vision-action AI model development, including $400 million for Generalist AI Inc., a startup working on embodied robotics intelligence, and the Jeff Bezos-backed robotics software startup Physical Intelligence, which raised $600 million.

How these machines perceive the world is the top of the stack that feeds the intelligence layers.

With the new funding on board, Luxonis said it intends to drive further development of its edge AI architectures, including advancing the OAK4 ecosystem, and to expand its supply chain capacity. It also said it intends to grow its research and development, go-to-market and engineering support teams.

The company added that it plans to launch new devices with accessible price points and in flexible form factors. It intends to target a wide variety of industries that are currently adopting AI-driven robotics and intelligent automation, including defense, industrial and heavy machinery, medical technology and warehousing.

Image: Luxonis