Robots are part of an exciting new area in technology, but it’s a challenge: Robots rely on sensors, external signals such as GPS and Wi-Fi, and customized software to browse their environment. Furthermore, robotics often involve expensive off-the-shelf hardware solutions, including built-in software and sensors for specific tasks, such as estimating relative motion. These products require complex integration and are limited to specific use cases.
As a result, most robots today cannot move between different locations, and only a small number of autonomous driving systems use AI to navigate.
But Tony Zhang, founder and CEO of Tera AI, believes that the software is called Zero shot The robot’s navigation can overcome these obstacles – the investor just gave him $7.8 million in seed funding to prove it.
High level tera ai A Spatial reasoning AI Systems that provide affordable visual navigation for automated robots. The technology is used in a variety of applications, including robotic manipulation, mobile robotics technology and autonomous driving.
“We play pure software, platform-inclusive approach to using any robot with pre-existing cameras and GPUs,” Zhang said in an interview with TechCrunch. “The system is cognitively inspired and can be applied to completely novel scenarios in reasoning time – a bit like a large language model (LLM).
Zhang founded San Francisco-based TERA in 2023 after leading machine learning at Google X, where he developed and commercialized geospatial models. He received his PhD from Caltech Pietro Peronaa pioneer in computer vision, examines how biological systems solve navigation in a general way.
The startup’s team includes AI and simulation researchers from Google AI, Caltech, MIT and European Space Agency.
Although much of the AI industry is focused on LLM, Zhang and his team have developed a new approach that enables AI to learn spatial reasoning independently. Spatial reasoning AI allows machines to navigate, identify objects and interact with three-dimensional space. Zhang told TechCrunch that universal navigation software that eliminates hardware constraints can greatly reduce costs and implementation time, thereby increasing the robot by 1,000 times.
“This could also allow existing robots to provide new capabilities for existing robots in areas where they simply cannot be autonomous due to restrictions on sensors,” he said.
For example, a $250,000 Waymo vehicle can afford a $50,000 localized sensor and a $100,000 laser system. However, the lighter robot costs under $50,000 and requires more affordable solutions to automatically navigate. Additionally, the cost of a high-precision GPS receiver can be $10,000, while the top IMU (Inertial Measurement Unit) can reach $30,000, which makes it impossible for many smaller robots to achieve the cost of automatic navigation.
“Our key unique value proposition is that we are totally hardware agnostic, which means we are focused on solving universal navigation in pure software in any robot and any new environment without every readjustment,” Zhang said. “This is the first time in robotics, we can sell a software that works like an operating system, allowing any mobile robot platform to reach its full potential and deliver on its promise.”
The startup has been testing its products with a variety of major U.S.-based players in the robotics industry. The company’s customers are primarily robot manufacturers, who already have customers but face challenges in expanding solutions to different autonomous platforms, situations and environments.
New funding will help TERA deploy its initial solutions on embedded devices this year and expand its technical team.
Zhang told TechCrunch: “Once people realize that existing cameras already on robots are enough to locate and navigate, the software will become the most valuable robot platform. “Ultimately, we envision a future where, like the iOS App Store, you just click to download and thrive to install new features – your robot has a whole new capability. ”
Investors in the Tera seed round include Felicis, Inovia, Caltech, Wilson Hill and entrepreneur Ravikant.
