RoboBoom
Humanoids Summit· July 7, 2025· 17:05

Santiago Pavon (Pollen Robotics) | Open-Source Embodied AI & Reachy 2 | Humanoids Summit London 2025

Santiago Pavon, growth lead at Pollen Robotics, discusses open-source robotics, Reachy 2, and ways to make embodied AI more accessible with the Humanoids Summit host in London.

In this talk

  1. 0:34

    Open source as a way to bring people in

    Pavon says Pollen Robotics joined Hugging Face on April 15, while keeping its identity and independent way of working. He sees a natural fit in their shared commitment to open science and open-source tools, and credits LeRobot with helping researchers channel their interest into collaboration.

  2. 2:08

    Lowering the barriers to experimentation

    Pavon argues robotics needs more participation from people and regions that have had less access to the field. He suggests companies offer hands-on entry points, such as software demonstrations, simulations, inexpensive developer kits, and 3D-printable parts, so people can learn before investing in a full system.

  3. 4:23

    Community can create lasting value

    Pavon says open technologies let users adapt their tools and choose what works, rather than depend on one company or business model. He also sees value in community engagement and product adoption alongside hardware and software sales, including the activity and collaboration a platform attracts.

  4. 9:27

    Expanding access while keeping performance

    Pavon acknowledges Reachy is a high-end robot, priced around $70,000, and says its specifications serve clients seeking a capable platform. Pollen is exploring ways to let people try its technology without buying a robot, and he says a more affordable Reachy could arrive in 2026 or 2027. The team is also presenting open-source, 3D-printable projects costing under €300.

  5. 13:39

    Iterating with the community

    Pavon says Pollen has four to seven projects in progress and expects a range of smaller robotics projects alongside Reachy 2 in 2026. The team watches community response, improves early versions with user feedback, and moves on when a project lacks traction. These projects may cover areas such as manipulation, computer vision, machine learning, and 3D printing.

Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.

More talks