On the 7th and 8th of May 2026, our Principal Research Scientist, Pedro Ortiz Suarez, had the pleasure of participating in the founding workshop of Project Tapestry. This project, whose goal is to develop open-source and sovereign AI at a global scale, was launched in Paris during this event that gathered around 50 researchers, founders and engineers from public and private institutions around the world.
Project Tapestry, which is led by the AI Alliance and Turing Award winner Yann LeCun, has as a goal to create a federated platform of compute and foundational models that would allow institutions, industries and nations to develop derivatives of a base model that are aligned with their culture, values, laws and priorities.
The event took place over two days, with the first day featuring talks from Yann LeCun and the AI Alliance introducing the project, as well as presentations from existing initiatives of open and sovereign AI efforts. These included talks from Antoine Bosselut (Apertus Model), Hector Liu (MBZUAI - K2 Initiative) and Ayah Bdeir (Current AI). These were followed by technical discussions on Project Tapestry and how to scale the presented initiatives to a global, multilingual and multicultural effort. The day ended with a dinner where participants had the chance to exchange ideas in a more informal and relaxed setting.
The second day included a couple of technical talks from Ganesh Ramakrishnan presenting BharatGen, as from Michitaka Tsuda on Japan’s government work on Open Data Spaces. These presentations were followed by technical and in-depth discussions on how to start working together towards the goals of Project Tapestry. The second day ended with dinner at a Parisian bistrot.
The Common Crawl Foundation is happy to be part of this initiative, as enabling culturally-informed and sovereign technologies has been at the core of our mission for 18 years. We believe that Project Tapestry faces a big challenge in terms of responsibly and politely managing data to ultimately develop this global framework for sovereign models, we believe that data is at the core of this effort and as such, we would like to offer our experience in engaging with linguistic and cultural communities around the world in recent Common Crawl projects aiming to expand the linguistic and cultural coverage of our dataset, such as CommonLID and the Web Languages Project. Efforts that Pedro presently presented at the IBM headquarters in New York City, during the UN Open Source Week.
We look forward to a productive collaboration with Project Tapestry and with organizations involved in this project.
.svg)
