Hugging Face

• 1. Founding And Pivot

  ›

  Clément Delangue, Julien Chaumond, and Thomas Wolf founded Hugging Face in 2016. They established the company in New York City with a specific goal. The original product was a chatbot application designed for teenagers. This early version took its name from the popular emoji symbol used to express affection. The founders released the underlying model as open source code shortly after launch. That decision changed the trajectory of the business entirely. The team shifted focus away from consumer apps toward infrastructure for machine learning developers. They built tools that allowed researchers to share models and datasets freely.

  2

• 2. BigScience And BLOOM

  ›

  The BigScience Research Workshop launched on the 28th of April 2021. Multiple research groups collaborated to create an open large language model. The initiative concluded in 2022 with the release of BLOOM. This multilingual model contained 176 billion parameters. It represented a significant step forward in natural language processing capabilities. The project demonstrated how distributed teams could build complex AI systems together. Researchers across the globe contributed to the training process. The resulting tool supported dozens of languages simultaneously. It became a cornerstone for subsequent open-source developments in the field.

  345

• 3. Strategic Partnerships

  ›

  Amazon Web Services announced a partnership with Hugging Face in February 2023. AWS customers gained access to Hugging Face products as building blocks for custom applications. The collaboration included plans to run future versions of BLOOM on Trainium chips. These proprietary machine learning chips were developed by Amazon specifically for this purpose. Thomas Wolf led discussions regarding technical integration during early 2024 meetings. Scaleway also joined the ecosystem to provide additional computing power. These alliances expanded infrastructure availability for developers worldwide. They reduced barriers to entry for smaller organizations seeking advanced tools.

  678

• 4. European AI Initiatives

  ›

  A new accelerator program launched in June 2024 involving Meta and Scaleway. The initiative targeted European startups aiming to integrate open foundation models into their products. STATION F in Paris served as the physical base for the program. Participants received mentoring alongside access to AI models and computational resources. The cohort ran from September 2024 through February 2025. Selected teams gained direct support from industry experts. This effort aimed to strengthen the EU AI ecosystem significantly. It provided critical resources for emerging companies developing localized solutions.

  9

• 5. Indigenous Language Preservation

  ›

  On the 23rd of September 2024, Hugging Face partnered with UNESCO and Meta. Their goal was to launch an online language translator supporting over 200 languages. The tool utilized Meta's No Language Left Behind open-source AI model. It enabled free text translation across many low-resource languages. This project aligned with the International Decade of Indigenous Languages. The collaboration sought to preserve linguistic diversity through technology. Developers could now translate content between previously unsupported dialects easily. The system prioritized accessibility for communities lacking digital representation.

  1011

• 6. Robotics Expansion

  ›

  Hugging Face acquired Pollen Robotics in April 2025. The French humanoid robotics startup had been founded by Matthieu Lapeyre and Pierre Rouanet in 2016. Clément Delangue announced the acquisition via a post on X. He stated his vision to make artificial intelligence robotics open source. This move signaled a shift toward physical computing applications. The company planned to integrate robotic systems with its existing platform. Developers would gain access to shared models for autonomous machines. The acquisition expanded the scope beyond software into tangible hardware interactions.

  121314