The AI Alliance - the nonprofit consortium founded by IBM and Meta - published its first planning report for Project Tapestry, an initiative testing whether frontier-scale AI development can be distributed across a global network of institutions rather than concentrated inside a single lab.
Around 30 researchers and institutional representatives met in Paris in May. Participating organizations include Switzerland's Apertus, India's BharatGen, the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) from the UAE, and AI Singapore. Yann LeCun, Meta's chief AI scientist and one of the more vocal critics of closed-lab AI development, is serving as chief science advisor.
The premise runs counter to how frontier AI has actually been built. OpenAI, Anthropic, Google DeepMind, and xAI have all taken the centralized, heavily-funded private lab route. Project Tapestry is asking whether international institutions can pool compute, data, and research talent to build something competitive without any single company controlling the result. LeCun has argued for years that distributed, open development is the better path - Project Tapestry gives him a formal vehicle to test that argument.
This is still early. A planning workshop is a long way from a working frontier model, and participating organizations would need to coordinate across countries with different data laws, compute access, and regulatory frameworks. But the institutional backing is real, and the question Project Tapestry is trying to answer matters: the current trajectory concentrates enormous power in a handful of US-based companies. Whether a coalition model can actually compete is what this initiative exists to find out.