In a move that signals a deepening commitment to physical AI, Hugging Face has acquired Pollen Robotics, a French humanoid robotics startup. The transaction, whose financial terms were not disclosed, brings the Bordeaux-based company under the umbrella of the AI development platform best known for its open-source model repository.
The acquisition was reported in April 2025, according to coverage from TechCrunch. Under the arrangement, Hugging Face intends to commercialise Pollen’s humanoid robot, Reachy 2, and will allow developers to download the robot’s code and propose improvements to it. This open-software approach is consistent with Hugging Face’s broader philosophy of making AI tools accessible to a wide community of developers.
Pollen Robotics was founded in 2016 by Matthieu Lapeyre and Pierre Rouanet. The company’s stated mission has been to bring affordable humanoid robots into the home — a goal that has proven elusive for many robotics firms, given the high cost of actuators, sensors, and manufacturing. Prior to the acquisition, Pollen had raised approximately €2.5 million (around $2.83 million) from investors, including Bpifrance, according to data from Crunchbase cited in the TechCrunch report. That funding figure is modest by robotics industry standards, where hardware startups often require tens of millions of euros to reach production scale.
The acquisition is not Hugging Face’s first foray into robotics. The company launched LeRobot in 2024, a collection of open AI models, datasets, and tools designed for building robotics systems. LeRobot has become a central pillar of Hugging Face’s robotics strategy, and the Pollen team is expected to contribute significantly to this effort.
Hugging Face has also established a dedicated robotics team led by Remi Cadene, a former robotics engineer who previously worked on Tesla’s Optimus humanoid program. Cadene’s background in one of the most high-profile humanoid robot projects in the world gives Hugging Face’s robotics division a degree of credibility in a field that is crowded with ambitious startups and established industrial players alike.
The acquisition follows a period of collaboration between the two companies. Last year, Hugging Face teamed up with Pollen to build “Le Robot,” an open-source robot trained to perform a variety of household chores. That project demonstrated the potential for combining Hugging Face’s AI models with Pollen’s hardware expertise, and it appears to have laid the groundwork for the full acquisition.
Why it matters for European robot service
The acquisition of Pollen Robotics by Hugging Face is significant for the European robotics ecosystem for several reasons. First, it represents a validation of French robotics talent. Pollen Robotics emerged from the French startup ecosystem with a relatively small amount of funding — €2.5 million is a fraction of what many robotics companies raise in their seed rounds. Despite this constraint, the company managed to develop a humanoid robot that attracted the attention of one of the most prominent AI companies in the world.
Second, the deal highlights the growing convergence of AI software and physical robotics. Hugging Face is primarily known as a software company — its platform hosts hundreds of thousands of machine learning models that developers use for natural language processing, computer vision, and other AI tasks. By acquiring Pollen, Hugging Face is making a clear statement that it intends to be a player in the physical world, not just the digital one.
For European robot service providers, this acquisition could have several implications. The most immediate is the availability of Reachy 2 as a commercial product. Hugging Face has stated that it plans to sell the robot, which means that European companies and research institutions may have a new option for acquiring a humanoid platform. The fact that developers will be able to download and improve the robot’s code is particularly notable — it suggests that Reachy 2 could become a platform for experimentation and customisation, rather than a closed, proprietary system.
The acquisition also signals that Hugging Face is serious about building out its robotics capabilities. The company has been making a concerted push into the robotics industry over the past few years, and the Pollen acquisition is the most significant step yet. In 2025, Hugging Face released an updated version of its 3D-printed and programmable robotic arm, the SO-101, which the company built in partnership with French robotics firm The Robot Studio. This arm is the follow-up to the SO-100, which was released the previous year for around $100 — a price point that made it accessible to hobbyists, educators, and researchers.
Hugging Face has also expanded the training data on its LeRobot platform through a partnership with AI startup Yaak. This partnership added training data for self-driving machines, broadening the platform’s applicability beyond stationary or semi-mobile robots to include autonomous vehicles. This expansion suggests that Hugging Face is thinking about robotics in a comprehensive way, covering everything from desktop robotic arms to humanoid robots to self-driving machines.
For European robot service companies, the key takeaway is that the landscape is shifting. The traditional boundaries between AI software companies and hardware manufacturers are blurring. Hugging Face, a company that many in the robotics industry might have viewed as a pure software player, is now selling humanoid robots and robotic arms. This could create new opportunities for collaboration, but it could also introduce a new competitor into a market that is already crowded.
The European robotics market has long been dominated by industrial players like ABB, KUKA, and Universal Robots, which focus on factory automation. In recent years, a new wave of startups has emerged, focusing on service robots, collaborative robots, and humanoid robots. Pollen Robotics was part of this new wave, and its acquisition by Hugging Face could serve as a template for other European robotics startups looking for an exit. Rather than going public or being acquired by a traditional industrial conglomerate, these startups may find that AI companies are increasingly interested in acquiring hardware expertise.
What buyers and operators should know
For buyers and operators considering Reachy 2 or other Hugging Face robotics products, there are several important considerations to keep in mind.
First, the financial details of the acquisition have not been disclosed. This means that it is unclear how much Hugging Face paid for Pollen, and what the financial health of Pollen was at the time of the acquisition. Pollen had raised €2.5 million prior to its exit, according to Crunchbase, which is a relatively small amount for a hardware company. This could mean that the company was operating on a lean budget, or it could mean that the company had difficulty raising additional funding. Without more information, it is difficult to assess the financial trajectory of the company.
Second, the availability of Reachy 2 as a commercial product is still being established. Hugging Face has stated that it plans to sell the robot, but the specifics of pricing, delivery timelines, and after-sales support have not been detailed in the source material. Buyers should be aware that these details may not be fully resolved, and they should seek clarification from Hugging Face before making any purchasing decisions.
Third, the open-source nature of Reachy 2’s code is both an opportunity and a consideration. On one hand, the ability to download and modify the code means that developers can customise the robot to their specific needs. This could be particularly valuable for research institutions and companies that need to integrate the robot into existing workflows. On the other hand, open-source hardware and software can come with challenges. There may be limited official support, and buyers may need to rely on community forums and documentation rather than a dedicated customer service team.
Fourth, it is worth noting that Hugging Face’s robotics division is led by Remi Cadene, who previously worked on Tesla’s Optimus program. This gives the team a degree of credibility, but it also means that the team may have a particular vision for humanoid robotics that is informed by the Tesla approach. Buyers should consider whether this vision aligns with their own needs.
Fifth, the broader context of Hugging Face’s robotics push is important. The company has released the SO-101 robotic arm, which is a 3D-printed, programmable arm that can pick up and place objects and perform a few other basic chores. This arm was built in partnership with The Robot Studio, and it is the follow-up to the SO-100, which was released for around $100. The low price point of the SO-100 made it accessible to a wide range of users, and the SO-101 appears to be continuing this trend. For buyers who are new to robotics, the SO-101 could be a more accessible entry point than a full humanoid robot like Reachy 2.
Sixth, the partnership with Yaak to expand LeRobot’s training data to include self-driving machines is a sign that Hugging Face is thinking about the full spectrum of robotics applications. This could mean that the company’s robotics platform will eventually support a wide range of robots, from stationary arms to mobile platforms to autonomous vehicles. Buyers who are considering investing in Hugging Face’s robotics ecosystem should be aware that the platform is likely to evolve and expand over time.
Finally, it is important to note that the acquisition of Pollen Robotics is part of a larger trend of AI companies moving into hardware. Hugging Face is not the only AI company to make this move, and it is unlikely to be the last. For buyers and operators, this trend could be positive, as it may lead to more innovative products and more competitive pricing. However, it could also introduce uncertainty, as companies navigate the challenges of hardware manufacturing, supply chain management, and customer support.
One thing that is not disclosed in the source material is the specific timeline for the availability of Reachy 2 under Hugging Face’s ownership. The acquisition was announced in April 2025, but it is unclear when the robot will be available for purchase, what the pricing will be, or what the delivery lead times will be. Buyers should not assume that Reachy 2 is immediately available, and they should contact Hugging Face directly for the most current information.
Similarly, the source material does not specify what level of support Hugging Face will provide for Reachy 2. Will there be a dedicated support team? Will there be warranty coverage? Will spare parts be readily available? These are important questions for any buyer considering a humanoid robot, and they are not answered in the source material. Buyers should seek clarity on these issues before making a commitment.
It is also worth noting that the source material does not provide any information about the performance specifications of Reachy 2. The robot is described as a humanoid robot, and the source material mentions that it was involved in the “Le Robot” project, which was trained to do household chores. However, there is no information about the robot’s payload capacity, battery life, processing power, or other technical specifications. Buyers who need this information will need to seek it from Hugging Face or from other sources.
In summary, the acquisition of Pollen Robotics by Hugging Face is a significant development in the European robotics landscape. It brings a French humanoid robotics startup under the umbrella of one of the most prominent AI companies in the world, and it signals a deepening commitment to physical AI. For buyers and operators, the key considerations are the availability of Reachy 2, the open-source nature of its code, the broader context of Hugging Face’s robotics push, and the many details that have not yet been disclosed. As with any emerging technology, it is wise to approach with informed caution and to seek out the most current information before making any decisions.
Sources
Published by Vigla Media OÜ (Estonia).