New Technology / Robotics
Track robotics trends, industrial automation, machine intelligence and commercial deployment signals through curated technology summaries.
Figure 03 Reveals 2026 Humanoid Robot Endgame (AI NEWS)
Topic
Advancements in Humanoid Robotics
Key insights
- Figures CEO announced plans to deploy robots on production lines this year, aiming for full robot-built robots within 24 months and predicting billions of humanoids in commercial use
- Figure's CEO announced plans to deploy robots on production lines this year, aiming for full robot-built robots within 24 months. The company predicts billions of humanoids in commercial use, indicating a significant shift in automation and robotics.
- Dream Dojo, trained on 44,000 hours of human egocentric videos, enhances humanoid robots understanding of complex physical interactions and enables realistic action simulations
- It generates video rollouts for robots like GR1 and G1, improving generalization to new objects and environments
- Achieving real-time performance at 10 to 11 frames per second, Dream Dojo supports long-horizon simulations crucial for various applications
- The system excels in handling precise, contact-rich interactions, essential for tasks requiring high dexterity
Perspectives
Focus on advancements and implications in humanoid robotics.
Figure's CEO
- Announces plans to deploy robots on production lines this year
- Aims for full robot-built robots within 24 months
- Predicts billions of humanoids in commercial use
- Highlights superhuman speed and agility advancements
- Foresees robots performing surgery by the end of the year
- Envisions humanoids transforming workplaces and space exploration
Unitree and NVIDIA Researchers
- Demonstrate G1 humanoid performing swarm acrobatics
- Question the authenticity of robot demonstrations
- Unveil Dream Dojo for simulating complex physical interactions
- Achieve fast real-time generation for humanoid robots
- Highlight the need for broad human-like physical intuition in robots
Neutral / Shared
- Discusses the potential for humanoids in various tasks
- Mentions the importance of rigorous testing and validation
Metrics
duration
15 times longer duration times
comparison to previous datasets
This indicates a significant increase in training data, enhancing model performance.
first person human video, which is 15 times longer duration
skills
96 times more skills times
comparison to previous datasets
A broader skill set allows for more complex task execution.
96 times more skills
scenes
2,000 times more scenes times
comparison to previous datasets
This diversity in scenes enhances the model's adaptability.
2,000 times more scenes than the previous biggest data set
real-time performance
10 to 11 frames per second fps
performance metric of Dream Dojo
This performance is crucial for smooth simulations and interactions.
Dream Dojo achieves fast real-time generation at 10 to 11 frames per second
Key entities
Timeline highlights
00:00–05:00
Figure's CEO announced plans to deploy robots on production lines this year, aiming for full robot-built robots within 24 months. The company predicts billions of humanoids in commercial use, indicating a significant shift in automation and robotics.
- Figures CEO announced plans to deploy robots on production lines this year, aiming for full robot-built robots within 24 months and predicting billions of humanoids in commercial use
05:00–10:00
Dream Dojo enhances humanoid robots' understanding of complex physical interactions through extensive training on human egocentric videos. It achieves real-time performance and supports various applications, including teleoperation and model-based planning.
- Dream Dojo, trained on 44,000 hours of human egocentric videos, enhances humanoid robots understanding of complex physical interactions and enables realistic action simulations
- It generates video rollouts for robots like GR1 and G1, improving generalization to new objects and environments
- Achieving real-time performance at 10 to 11 frames per second, Dream Dojo supports long-horizon simulations crucial for various applications
- The system excels in handling precise, contact-rich interactions, essential for tasks requiring high dexterity
- Downstream applications include teleoperation, revolutionizing human-robot interaction in diverse environments
- Dream Dojo enhances model-based planning and policy evaluation, vital for developing effective robotic systems