The robotics startup Skild AI receives financial support from Amazon and SoftBank to develop a general-purpose artificial intelligence model which operates across various robotic systems including humanoids and warehouse machines. The Skild Brain system provides robots with human-like capabilities through its ability to reason and adapt while making decisions in real time.
The model trains through a combination of human-action video data and simulated environments and real-time feedback from operational robots. Through continuous data loop processing the AI system enhances its abilities across different operational settings. The Skild Brain system demonstrated robotic capabilities through its ability to perform stair climbing and shove recovery and navigate through disorganized spaces which traditional automation systems find challenging.
The safety features of Skild include force-limiting capabilities which minimize potential harm to users. The company serves LG CNS and logistics organizations which utilize Skild systems for manufacturing and supply chain operations.
The model developed by co-founders Deepak Pathak and Abhinav Gupta who come from Carnegie Mellon and Meta addresses the fundamental robotics challenge of limited available data. According to Pathak there exists no extensive robotic data available on the internet. “We’re building it from the ground up.”
The company received $300 million in Series A funding during the previous year which elevated its value to $1.5 billion. The company has recruited personnel from Tesla Meta and Nvidia to become a leading force in developing industrial-grade intelligent robots for commercial markets.
