Skild AI has introduced S1, a robot foundation model designed to learn and carry out previously unseen tasks after watching a single video demonstration. According to the company, the model can handle new tasks lasting up to 10 minutes without updating model weights or applying task-specific post-training. Skild AI said the technology has already led to more than 60 deployment partnerships across manufacturing, logistics, and food preparation. Co-founder and CEO Deepak Pathak said learning from experience rather than pre-programming represents a major shift in robotics. The company also said NVIDIA Isaac Lab and Cosmos helped generate the large-scale and diverse experience needed for robot learning. In factory settings, Skild has worked with NVIDIA and Foxconn on high-precision assembly for NVIDIA Blackwell systems. Citing NVIDIA Blog, the report added that Skild AI reached a $100 million annual recurring revenue run rate 10 months after its first commercial deployment. In testing, S1 posted about a 66% per-step success rate on new multi-step tasks, compared with 9% for similar AI systems.
Skild AI has released its S1 robot foundation model, which the company said can learn and execute previously unseen tasks lasting up to 10 minutes by watching a single video demonstration. The model does not require weight updates or task-specific post-training.
Skild AI said the technology has formed more than 60 deployment partnerships across manufacturing, logistics, and food preparation.
Deepak Pathak, co-founder and CEO of Skild AI, said learning through experience instead of pre-programming is a transformative step for robotics. The company added that NVIDIA Isaac Lab and Cosmos helped Skild build the large-scale and diverse experience required for robot learning.
In factory environments, the model has been used with NVIDIA and Foxconn for high-precision assembly of NVIDIA Blackwell systems.
According to NVIDIA Blog, Skild AI reached a $100 million annual recurring revenue run rate 10 months after its first commercial deployment. In tests, the S1 model achieved an approximately 66% success rate at each step of new multi-step tasks, compared with 9% for similar AI systems, a gain of more than sevenfold.
This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan. Disclaimer:
The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.
Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.
