×
Hugging Face brings open-source revolution to humanoid robotics with Pollen acquisition
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

Hugging Face‘s acquisition of Pollen Robotics marks a significant step toward democratizing humanoid robotics through open-source development. By purchasing the company behind the two-armed Reachy 2 robot, Hugging Face is extending the open-source ethos that has accelerated AI progress into the physical robotics domain, potentially addressing the transparency challenges that have plagued recent humanoid robot demonstrations and development.

The big picture: Hugging Face plans to sell Pollen Robotics’ humanoid robot Reachy 2 while making its code openly available for developers to download, modify, and improve upon.

  • “It’s really important for robotics to be as open source as possible,” says Clément Delangue, Hugging Face CEO, emphasizing that physical robots demand greater trust and transparency than software-only AI systems.
  • The acquisition comes amid growing interest in combining advanced AI models with robotic hardware, with some researchers arguing that AI needs physical embodiment to reach human-level intelligence.

Why this matters: Open-source approaches have accelerated AI development across software and models, but robotics hardware has largely remained the domain of well-funded companies with proprietary systems.

  • Making humanoid robotics open-source could potentially accelerate innovation in the field similar to how open AI models like Meta‘s Llama have contributed to rapid progress in language models.
  • The approach directly challenges the closed development methodologies of companies like Tesla, Figure, and Agility Robotics that currently dominate the humanoid robotics space.

What they’re saying: Hugging Face and Pollen Robotics executives emphasize transparency as a central benefit of their open-source strategy.

  • “When you think about physical objects doing physical things at work and at home, the level of trust and transparency I need is much higher than for something I chat with on my laptop,” explains Delangue.
  • “You can’t cheat, you can’t hide with open source,” Delangue notes, addressing concerns about misleading robot demonstration videos from competitors.

Current capabilities: Reachy 2 has been shown performing basic manipulation tasks in demonstrations shared by Pollen Robotics.

  • The bug-eyed, two-armed humanoid robot can tidy coffee mugs and pick up fruit, according to videos from the company.
  • Several leading AI companies are already using Reachy 2 for robotic manipulation research, according to Matthieu Lapeyre, Pollen Robotics’ cofounder and CEO.

Industry challenges: The humanoid robotics sector faces significant hurdles despite recent investor enthusiasm.

  • Lapeyre acknowledges that selling humanoid robots remains challenging due to unclear use cases and reliability issues.
  • Experts caution that impressive-looking demonstration videos may be misleading, potentially showing teleoperated systems, single-attempt successes, or robots that can’t reliably complete shown tasks.

The industry trend: Hugging Face’s move aligns with a broader shift toward more open approaches in artificial intelligence development.

  • Meta pioneered the release of cutting-edge open weight models with Llama in 2023.
  • Even OpenAI, known for its closed approach, announced plans to release a free, open weight model in summer 2024.

Expert perspective: Academic researchers see tangible benefits in open-source robotics development.

  • “Making robotics more accessible increases the velocity with which technology advances,” says Sergey Levine, a UC Berkeley assistant professor and cofounder of robotics startup Physical Intelligence.
An Open Source Pioneer Wants to Unleash Open Source AI Robots

Recent News

MILS AI model sees and hears without training, GitHub code released

Meta researchers develop system enabling language models to process images and audio without specialized training, leveraging existing capabilities through an innovative inference method.

Mayo Clinic combats AI hallucinations with “reverse RAG” technique

Mayo's innovative verification system traces each AI-generated medical fact back to its source, dramatically reducing hallucinations in clinical applications while maintaining healthcare's rigorous accuracy standards.

Columbia dropouts launch Cluely, an AI tool designed for cheating in interviews and exams

Columbia dropouts' desktop AI assistant provides real-time answers during interviews and exams through an overlay invisible during screen sharing.