Project Tapestry
Welcome to The AI Alliance Project Tapestry.
The AI Alliance launched Project Tapestry to build a collaborative foundation for open and sovereign AI. Project Tapestry will be an open-source platform designed to enable globally federated development of frontier open models while preserving sovereignty, local control, and long-term independence.
Turing Award winner Yann LeCun is the Chief Science Advisor to the AI Alliance and Project Tapestry.
- Join Us! We are looking for collaborators. See our contributing page for details.
- Use the search box at the top of this page to find specific content.
- The links for Capitalized Terms go to this glossary.
Why Project Tapestry?
Today, open-weight AI models are everywhere. But open weights alone do not make pretraining participatory. The infrastructure, data pipelines, and design decisions behind these models remain concentrated in a small number of companies and regions. Most of the world downloads the resulting models, but almost no one shapes the process of building them. Yann LeCun calls this the central challenge of the next era of AI development.
Project Tapestry offers a new path for advanced AI development, built on advances in distributed training, which have demonstrated that globally federated model development can match synchronous baselines. Project Tapestry enables institutions, industries, and nations to co-train a shared, open-foundation model, while retaining control of their data and the ability to build sovereign derivatives aligned to their own priorities and requirements.
With Project Tapestry, the AI Alliance aims to create a new path for advanced AI development, one in which institutions, industries, and nations can join together to contribute to build more capable and open base models, while retaining control of their data and the ability to build on the base models to produce sovereign, derivative models aligned to their own priorities, industries, culture, laws, and values.
Frontier Open Models without Centralizing Data and Compute
Today, the development of the most capable AI models is increasingly concentrated within a small number of companies and regions. Open-weight models have expanded access, but core decisions about training data, model objectives, architecture, and evaluation typically remain concentrated within the institutions that originate them. At the same time, many sovereign AI model efforts face steep barriers in compute, funding, data access, and specialized talent.
Project Tapestry offers a alternative: a collaborative approach to AI model development in which participants can help build a shared open foundation without surrendering their data, strategic autonomy, or downstream control. Project Tapestry’s long-term vision is to develop an open, global model family, a shared open source base foundation model family that can draw on broader pools of expertise, compute, and domain knowledge than any single organization can typically assemble alone, while enabling participants to create sovereign derivative models tailored to their own societal, industrial, scientific, or mission-specific needs, and aligned to their own governance frameworks, languages, values, and priorities.
From Launch to Global Technical Mobilization
Over time, the AI Alliance expects Project Tapestry to foster a vibrant and enduring ecosystem of collaborative, sovereignty-preserving model building and application development. The AI Alliance’s 501(c)(3) nonprofit research organization will serve as the community home for Project Tapestry. It will host and support the Project Tapestry platform and open source technical assets including models that the AI Alliance develops with the platform. Project Tapestry will be governed by a board of representatives from major contributing organizations globally.
Watch for announcements over the coming months about Project Tapestry’s architecture, roadmap, and model development priorities.
Additional links
Some additional links.
- Contributing: We welcome your contributions! Here’s how you can contribute.
- About Us: More about the AI Alliance and this project.
- Project GitHub Repo
- The AI Alliance: The AI Alliance website.