Natura-1: 50B-param MoE robot foundation model pretrained on 1M hours of egocentric human video
heyshrutimishra · x · 2026-10-13
naturewill unveiled Natura-1, its first general foundation model for the physical world: a 50-billion-parameter mixture-of-experts model pretrained on 1 million hours of egocentric footage of human hands doing real work.
- Follows complex natural-language instructions zero-shot, reasoning over spatial constraints, multi-object relationships, and negative conditions
- Efficient post-training on limited robot data adapts pretrained world knowledge to dexterous manipulation
- Generalizes to unseen scenarios: given an outcome, it holds the goal while changing how it gets there
A stark contrast to factory arms that repeat one programmed job forever.
Related event: Nature Will Releases Natura-1, a 50B-Parameter Robotics Foundation Model(4 posts)→
More from Embodied
- Collov Labs launches NewEyes AI, giving personal agents continuous sight and memory — Div_pradeep · 2026-10-13
- Sipeed teases MaixStation powered by AXEROA AX8860: 256TOPS, 128GB LPDDR5X, aiming at DGX Spark — Marcuss2 · 2026-10-13
- Alibaba's Qwen launches AI glasses N1 and clip-on earbuds, glasses from 2,218 yuan — 千问APP · 2026-10-13
- Companion robot now befriends 1,500+ seniors, cutting agitation by 74% — CurieuxExplorer · 2026-10-13
- Causality researcher launches Aether AI, unveils CRIS-0 robotics system built on causal reasoning — s_mohinii · 2026-10-13
- Robot baristas make good coffee, but can't give you the joy of making it — zijing_wu · 2026-10-13