A striking pattern in the 2025 Hugging Face repository data is just how strongly major tech companies are now contributing to the open-source AI ecosystem.
This marks a clear shift from only a few years ago, when only a handful of big players, in particular Meta with the Llama family, were releasing large-scale open models. Repositories can represent new model families, fine-tuned variants, training and inference pipelines, or evaluation datasets, all tangible building blocks for the global AI community.
Unlike the early internet era, when big companies initially focused on proprietary software before gradually embracing open source, today’s open-source AI is being driven from the start by global tech leaders who actively enable the community to build on top of their work. Google, Meta, Microsoft, Tencent, Alibaba, Baidu, and NVIDIA are all releasing models, tools, and research assets at a scale that is reshaping the ecosystem.
NVIDIA clearly stands out in this picture, reaching around 360 repositories by December. This shows a commitment to open-source AI that goes far beyond its hardware roots, a remarkable evolution in strategy, contribution, and community-building. Jensen Huang even highlighted an earlier version of this graph in his GTC keynote in October to highlight how foundational open models have become for the next wave of AI development.
What is also remarkable is the rise - and increasing maturity - of the Chinese open-source AI ecosystem this year. Alibaba’s Qwen series has become one of the world’s most widely adopted open models, but Baidu and Tencent are other heavyweight contributors.
Special note goes out to Allen AI (Ai2), the non-profit organization behind the Olmo series currently tops the Artificial Analysis' Openness Index and has been very active this year, adding over 236 repositories to Hugging Face.
The overall message is encouraging: 2025 is the year Big Tech became one of the most powerful engines of open-source AI progress, accelerating an ecosystem that benefits everyone. The next big question is: what does this mature phase of open source make possible?