BMW, Toyota, Osram invest in AI processor startup
Recogni, founded in 2018 and based in Munich and Silicon Valley, has set itself the goal of revolutionizing the processing of sensor data for (semi)-autonomous vehicles. The company aims to solve the dilemma of ensuring the environmental perception of autonomous vehicles at high speed (real-time) on the one hand and with very low power consumption on the other. The company’s founders have extensive industry experience in system design, artificial intelligence, machine vision and chip design.
Osrams’ investment is made via its venture capital unit Fluxunit. In addition to Osram / Fluxunit, a number of well-known venture capitalists are also involved. These include GreatPoint Ventures, Toyota AI Ventures, BMW iVentures, Faurecia and DNS Capital. The initial funding round is $25 million.
“We see the Recogni platform as a great opportunity to truly achieve the goal of full vehicle autonomy,” said Ashok Krishnamurthi, Managing Partner of GreatPoint Ventures. “Our market research shows that most technologies for accelerating neural networks are trimmed for either performance or power efficiency – neither is optimized for both. We believe the Recogni platform is several orders of magnitude better than anything we’ve seen before.”
“We firmly believe in sensor fusion based on camera, radar and lidar, but the computing requirements for these algorithms remain one of the critical bottlenecks in autonomous driving,” said Sebastian Stamm, Investment Manager at Fluxunit. “Recogni solves this problem with a very innovative approach”.
Recogni plans to use the fresh capital to develop the most powerful perception system to date. It fuses visual and lidar/radar data using state-of-the-art technology. At the same time, the company continues to expand its development team. Recogni is in discussions with several car manufacturers to provide them with its basic technology from modules to software.
“The challenges within the autonomy ecosystem of levels 2+, 3, 4 and 5 range from the acquisition or generation of training data for neural networks to their real-time execution. In addition, these vehicles require the performance of a data center and yet consume only the smallest amount of energy,” explains RK Anand, CEO of Recogni, “In developing an all-new system for the automotive industry with a focus on maximum efficiency and lowest energy consumption, we benefit from our technological background in machine learning, computer vision, chip and system design.
For Osram, investing in Recogni is a step towards transforming itself from a lighting manufacturer into a high-tech photonics company active in various future fields such as autonomous driving. Through its venture capital arm Fluxunit, Osram plans to develop and contribute comprehensive expertise in optical semiconductors for lighting and sensor applications in the automotive sector.
Related articles:
Osram to acquire Ring Automotive: broadens aftermarket business
Osram partners with GaN Systems for ultra-fast LiDAR driver
CEVA acquires sensor processing technology company Hillcrest Labs
Quanergy and Chery partner on autonomous vehicles and smart cities
Kalray: Deep learning for high-performance applications