Startup Archetype AI has introduced its first "physical AI" model called Newton, designed to interpret sensor data and combine it with natural language to answer questions about the real world.
Unlike large language models or other multimodal models trained only on text and images, Archetype AI develops AI models that better understand the physical world. The company's first large-scale model, Newton, was trained using data from various sensors such as accelerometers, gyroscopes, radars, cameras, microphones, and thermometers.
The model is designed to combine sensor data with natural language to answer questions about the environment in real-time. Archetype AI's goal is for the model to extract patterns and insights from the real world that are too complex or fast-moving for humans to comprehend.
"The biggest problems in the world are physical, not digital," says Ivan Poupyrev, CEO and co-founder of Archetype AI. As examples, he noted that Newton could monitor motion sensors in a package and report if the contents were damaged, or it could interpret data in a home or factory and explain what is happening in plain language.
Early customers of Archetype AI include Infineon, Volkswagen, and unnamed companies in the automotive, consumer electronics, construction, logistics, and retail industries, the startup said. It has raised $13 million in seed funding led by venture capital firm Venrock, with additional investors including the Amazon Industrial Innovation Fund and Hitachi Ventures.
The new funding will be used to expand the Archetype AI team and scale the platform for multiple use cases across various industries, according to the press release.
"Archetype AI" leadership includes former Google researchers
The company aims to use Newton to replace time-consuming and expensive specialized solutions. It predicts that the accelerated convergence of terrestrial, atmospheric, and space-based sensors, combined with machine learning and data networks, will contribute to a "trillion sensor economy."
Archetype AI's founding team includes former executives from Google's Advanced Technology and Projects (ATAP) group. CEO Ivan Poupyrev and co-founders Brandon Barbello, Leonardo Giusti, Jaime Lien, and Nicholas Gillian bring decades of experience in AI, machine learning, hardware, and software design from companies like Disney, Sony, Samsung, NASA, Stanford, and MIT.