Project Tapestry

AI Alliance Launches Project Tapestry to Build a Collaborative Foundation for Open and Sovereign AI

The AI Alliance, a non-profit AI research and open-source technology coalition with more than 200 member organizations,  introduces Project Tapestry, a new open-source platform for globally federated development of frontier AI models — preserving local control and long-term independence.

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 result. Almost no one shapes the process. Turing Award winner Yann LeCun, who joins the initiative as Chief Science Advisor, 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 — the project 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.

The thesis is testable and the science is ready. What remains is coordination.

AI sovereignty has a number of dimensions. Fundamentally it means a person, organization, or nation can own and control their AI, use it how they see fit, and retain the value they create with it. Tapestry aims to enable AI sovereignty through both our technical platform design and through clear and rigorous governance of the project. We aim to foster:

1. National Sovereignty — Training data even for the base model does not need to leave a partner node. Nodes retain operational control over infrastructure, legal compliance, and what data participates in training. Data locality is enforced by design, not by agreement. Any node can create a derivative model they fully own.
2. Cultural Sovereignty — Apply custom alignment and policy layers to sovereign derivatives via fine-tuning, RL, constitutional AI, or DPO. A partner's cultural values shape their final model. Those choices never leak into the shared global base.
3. Industrial Sovereignty — Enhance the value and differentiation of products and services. Build domain-specific derivative and adapter layers for medical, legal, industrial, scientific, or mission-specific applications without giving up hard-earned proprietary data and knowledge, while taking advantage of the shared base. 

Tapestry is assembling its first contributors — the researchers, systems engineers, compute providers, governments, and institutions who will build the first sovereign federated training run. The AI Alliance is convening a founding workshop in Paris on May 7–8, 2026, bringing together technical leaders from around the world to define the architecture, roadmap, and model development priorities.

We are looking for:

1. ML researchers working on distributed optimization, federated learning, or low-communication training
2. Systems engineers with experience in large-scale GPU cluster orchestration
3. Compute providers — cloud, sovereign cloud, or national HPC centers
4. Government and policy leaders responsible for national AI strategy
5. Universities and research labs with multilingual, domain-specific, or institutional datasets

Project Tapestry is not a finished system — it is an open technical and institutional effort to build a durable collaborative foundation for open and sovereign AI.