AI Alliance launches Project Tapestry for open and sovereign AI
AI Alliance launched Project Tapestry on April 7, an open platform for federated frontier model training. Yann LeCun was announced at the same time as Chief Science Advisor.
What happened
On April 7, the AI Alliance launched Project Tapestry, a new effort to support distributed and globally federated training of frontier open models. The goal is to let countries, industries, and institutions help build base models without giving up control of their data, priorities, or local sovereignty. At the same time, Yann LeCun was named Chief Science Advisor.
Why it matters
This is a direct response to frontier AI development concentrating inside a small number of US companies. If Project Tapestry gains traction, it could become a new collaboration model for open source AI with stronger regional control and less dependence on a few platform vendors.
My take
It is early, but strategically important. Many European and public-sector organizations want AI capability without being fully locked into OpenAI, Anthropic, Google, or Meta. Project Tapestry points to a future where sovereignty matters almost as much as raw model quality. CIOs in regulated environments should watch this closely.
