Cambridge University spinout Matta is developing "sentient factories" that utilize artificial intelligence (AI) to monitor, comprehend, and optimize manufacturing processes in real time. The company, founded in 2022 and based in Shoreditch, London, emerged from research conducted at Cambridge University’s Institute for Manufacturing.
Matta's plug-and-play platform employs advanced algorithms to learn from production data. It identifies defects, diagnoses operational issues, and enhances manufacturing performance in real time. Since its first commercial deployment last year, Matta has been implementing its system in approximately two factories per month, contributing to the production of various items, including waterproof coats, loudspeakers, and robot arms. The company has a waiting list of over 300 industrial clients.
Last month, Matta announced it had secured $14 million in funding, comprising venture capital investments and grants from Innovate UK and the Royal Academy of Engineering, to further its technology development.
Matta's Vision and Technology
Dr. Douglas Brion, Matta's founder and CEO, stated the company's objective is to build AI foundation models that understand manufacturing operations and potential failure points. The system integrates various data types, such as vision and process data, through large AI models to perform tasks like anomaly detection, measurement, error correction, and, eventually, prediction of future issues.
Brion indicated that the technology aims to address fundamental engineering challenges in manufacturing. He highlighted the industry's reliance on human expertise and stated that Matta uses AI to capture and scale this tacit knowledge, facilitating the design of functional products. He also noted a perceived gap between AI specialists and manufacturing professionals, which Matta aims to bridge.
Current Applications and Future Expansion
Initial deployments of Matta’s system primarily replace manual inspection processes, which can be prone to errors and time-consuming. The system involves installing cameras on production lines to collect and learn from real-time data, typically becoming operational within hours and adapting to specific production lines within days.
For example, Matta's technology has been used for inspecting high-speed bottling lines for a global drinks brand and for rapidly measuring speaker components for Bowers & Wilkins. The latter application reduced the measurement time for speakers from a minimum of 20 minutes to 10 seconds.
The company is expanding its capabilities beyond defect detection to include problem-solving, collaborating with original equipment manufacturers (OEMs) to enable machines to self-tune. An example includes a partnership with additive manufacturing firm Caracol, where Matta’s vision AI is integrated for closed-loop control, linking real-time inspection to automatic parameter adjustments on industrial printers and large-format robot additive manufacturing cells.
Future plans involve leveraging collected data from various factories to predict potential issues during the design phase of products. The ultimate goal, as outlined by Brion, is to create a fully integrated system where engineers can design a part and have it manufactured efficiently.