Luxonis

Luxonis is a technology company specializing in artificial intelligence and computer vision. The company aims to simplify real-time image processing and perception tasks in industrial applications by developing embedded machine learning systems and edge artificial intelligence solutions. Headquartered in the United States, the company operates in more than 110 countries worldwide and is used by over 25,000 customers.

Founding

Luxonis began in 2019 with a real-time AI-powered safety project called Commute Guardian, designed to protect cyclists and pedestrians. This project formed the foundation of the company’s open-source visual perception platform, DepthAI. In 2020, Luxonis introduced the OAK-1 and OAK-D cameras, which quickly gained strong support on Kickstarter. In 2021, the product lineup expanded with the OAK-D Lite model. In 2022, ruggedized devices with Power over Ethernet (PoE) capabilities were developed for industrial use, and in 2023, the cloud-based Luxonis Hub platform was launched to enable centralized device management and model deployment. In 2024, the next-generation edge AI devices of the OAK Series 4 were released, bringing human-like perception capabilities even closer.

Technological Infrastructure and Product Ecosystem

Luxonis’s core products include the OAK (OpenCV AI Kit) cameras, the DepthAI platform, and the Luxonis Hub software. DepthAI is an open-source software and library suite capable of real-time data processing using stereo depth perception and neural network models. The OAK cameras are equipped with features such as stereo imaging, sub-pixel accuracy depth mapping, active stereo projection, image post-processing, and dynamic calibration. These devices offer computational power of up to 48 TOPS INT8 or 12 TOPS FP16 and can run models such as YOLOv5, ResNet-50, and BiSeNet at high frame rates. Luxonis Hub is a platform that enables centralized management of all devices and models. Its API-based architecture allows programmatic execution of device connectivity, model deployment, training, and monitoring. The Hub offers scalable architecture suitable for small-scale projects as well as enterprise systems with thousands of devices.

Privacy Principles

Luxonis devices process all data locally on the device, eliminating cloud dependency and minimizing latency. This architecture enables uninterrupted operation in scenarios where network connectivity is weak or interruptions are unacceptable. Additionally, since the devices process sensitive data such as Personally Identifiable Information (PII) locally, they comply with regulations including the European Union’s General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA).

Applications

Luxonis solutions are used across numerous industries including agriculture, logistics, manufacturing, robotics, healthcare, automotive, retail, and food services. These systems perform tasks in application areas such as quality control, inventory management, robot navigation, patient monitoring, and traffic surveillance. Thanks to active stereo technology, high-accuracy depth measurement is possible even under low-light conditions and on low-texture surfaces.

Open Source and Developer Support

Luxonis supports developers in line with its open-source philosophy by providing the DepthAI API, example code, technical documentation, and community forums. Developers can integrate their own models with popular deep learning frameworks such as TensorFlow, PyTorch, and ONNX and run them on OAK devices.

Organizational Structure

Luxonis’s founders include Shawn McLaughlin and Martin Peterlin, with Bradley Dillon serving as CEO. The company’s management team consists of experts in artificial intelligence, computer vision, product development, and customer relations.

Future Vision

Luxonis aims to evolve vision systems into self-improving artificial intelligence systems capable of independently training and deploying their own models. To this end, the company plans to make edge-based systems smarter, more flexible, and more autonomous. Its long-term goal is to build the technological infrastructure that enables machines to achieve human-like perception and response capabilities.