Axis Robotics(@axisrobotics):This is a milestone for robotics. @DynaRobotics' Dyna-2 is a world-action model pre-trained on 1M+ hours of human video. It validates clear scaling laws and, for the first time, human-to-robot cross-embodiment transfer with zero robot data in pre-training. This aligns with our belief: noisy, in-the-wild data converges to robust policies at scale. We generate both egocentric and sim data at scale through our data engine. And we’d love to explore how our high-value, high-diversity datasets could help push this frontier even further.

2026.08.11 03:44

This is a milestone for robotics. @DynaRobotics' Dyna-2 is a world-action model pre-trained on 1M+ hours of human video. It validates clear scaling laws and, for the first time, human-to-robot cross-embodiment transfer with zero robot data in pre-training. This aligns with our belief: noisy, in-the-wild data converges to robust policies at scale. We generate both egocentric and sim data at scale through our data engine. And we’d love to explore how our high-value, high-diversity datasets could help push this frontier even further.

Dyna Robotics@DynaRobotics

2026.08.10 16:44

Today we are introducing Dyna-2, a world-action model pre-trained on one million hours of human video. At this scale, for the first time, we discovered several new scaling laws: • world-action models exhibit scaling law on human data across four orders of magnitude, from 1000 to 1,000,000 hours, • this human data scaling law implied a scaling law on never seen robot data, • both data and objective matter; world modeling and scaling on video data are essential for cross-embodiment scaling transfer to emerge 🧵