Pittsburgh-based startup Skild AI has launched S1, a robotics model capable of learning physical tasks by watching a single video of a human performing them, without fine-tuning or hardware-specific adjustments. The underlying Skild Brain technology uses a two-layer hierarchical architecture trained on trillions of simulated physics episodes and millions of human action videos. Robots running S1 achieved task completion rates between 60% and 80% within hours of initial data collection. Founded in May 2023 by Carnegie Mellon University researchers, Skild AI has raised 1.7 billion dollars in total funding, including a roughly 1.4 billion dollar investment led by SoftBank in January 2026, valuing the company at over 14 billion dollars.
Source: Read the original article
