Berlin-based automotive AI startup Teraki has developed a platform that promises to bring a more than tenfold increase in efficiency to components used in automotive electronics, a €350 billion industry. Founded in 2014, the startup has now raised €2 million from Hong Kong-based Horizon Ventures and American Family Ventures, bringing its total funding to €4.7 million.
Teraki’s software allows for the scaling of insurance, predictive maintenance, and autonomous driving applications by providing access to more qualitative data. When embedded in automotive electronic systems, the software enables hardware to process more than 10 times more data – without loss of information to train – and run machine learning methods of its customers.
By applying techniques originally developed for quantum computing, Teraki is able to condense data to as little as two percent of its original size. This means that with Teraki pre-processing, applications run more than 10 times faster than implementations based on neural networks. With respect to sensor fusion, Teraki drastically reduces energy consumption and heat production due to lower computational tasks, while still delivering the algorithm detection and prediction performance that are essential for advanced driver assistance systems (ADAS) and autonomous vehicles.
“Data driven insights will be key to innovation in the automotive and automotive insurance sectors, as a result, capturing highly accurate information from cars is the basis needed to drive to these insights,” said Katelyn Johnson, principal at American Family Ventures. “Teraki aids in acquiring this information with the highest efficiency and accuracy rates. We are excited to support the growth of the company and its path towards enabling better insurance applications.”
“Winning the support of these renowned and leading VCs is further validation of our vision to truly enable efficient edge computing and scalable AI applications in the automotive markets,” said Daniel Richart, Teraki’s CEO and co-founder. “With this investment we will be able to accelerate the time to market for our signed customer contracts and serve the growing customer funnel.”
The company has also announced the launch of its Teraki DevCenter—a cloud-based data training and prototyping environment that allows customers to train Teraki’s algorithms on their data. Data training is an essential step used to teach AI models or machine learning algorithms how to make data-driven predictions or decisions by building a mathematical model from input data.
Unique to the industry, the DevCenter automates this complex process and provides development teams with the opportunity to quickly train Teraki’s machine learning algorithms based on their own data and to evaluate exactly what performance advantages Teraki’s technology can provide.
“With the DevCenter we have automated data training tasks, allowing development teams to test our solution with their own data more quickly,” said Markus Kopf, Teraki’s co-founder and CTO. “Automating this entire process is complex and difficult. For our current and future customers, this makes it much easier to experience for themselves what Teraki technology can do in terms of edge processing and performance improvements that can lower their hardware and data communication costs, improve their applications and algorithms, and create new possibilities in the automotive systems of tomorrow.”