Matroid provides an easy-to-use studio for creating and deploying detectors (computer vision models) to search visual media for people, behavior, objects, and events without programming. The platform can search any live stream or recorded video and deliver real-time notifications, and it can be deployed in the cloud or on-premise. Customers span construction, manufacturing, security, media, retail and other industries, using Matroid to reduce manual video-review costs, improve safety, and aid regulatory compliance. Matroid positions its product as a software-defined sensor layer that addresses the complexity, corner cases, and training costs that have limited large-scale computer vision deployments. The company plans to use new funding to accelerate product development and expand go-to-market efforts in manufacturing, industrial IoT (IIOT), and video security. Financially, Matroid completed a $20 million Series B and now reports $33.5 million in total funding. Matroid provides a web platform for building custom video recognition detectors that customers train on their own videos and images to recognize objects and events in video streams. The company is commercializing its technology in the security and media markets, offering services such as brand and executive monitoring on TV and continuous CCTV anomaly detection. Matroid positions its approach as more democratized than traditional object-recognition APIs by enabling creation and sharing of detectors. Founder Reza Zadeh leverages his Stanford connections and the Scaled Machine Learning conference to build a community of researchers from Google, Intel, NVIDIA and OpenAI to support detector development. The company announced a $10 million Series A from NEA and Intel Capital, bringing total financing to $13.5 million. Rather than building its own hardware, Matroid is partnering with Intel to execute computer-vision tasks on Intel chips, with Intel acting as both a strategic investor and a customer.

Total raised

$30M

Funding rounds

2

Latest round

Series B

Latest activity

Oct 2020