Open Source and the future of European AI sovereignty: Insights from Vivatech 2025 – Open Source Initiative

Learn more about OSI’s mission, history, people, and programs.
Explore the list of OSI-Approved licenses and learn more about the license review process.
Learn more about Open Source AI, how do we define it, and its benefits to society.
Advancing Open Source through policy, regulation, and standards.
Supporting research, education, and knowledge about Open Source.
Connecting the people and organizations that build and support Open Source.
Explore the latest news, blog posts, newsletters, press mentions, and events from the OSI.
Learn more about OSI’s mission, history, people, and programs.
Explore the list of OSI-Approved licenses and learn more about the license review process.
Learn more about Open Source AI, how do we define it, and its benefits to society.
Advancing Open Source through policy, regulation, and standards.
Supporting research, education, and knowledge about Open Source.
Connecting the people and organizations that build and support Open Source.
Explore the latest news, blog posts, newsletters, press mentions, and events from the OSI.
French President Emmanuel Macron alongside NVIDIA CEO Jensen Huang and Mistral AI’s Arthur Mensch celebrated a landmark partnership at Vivatech 2025 last week aimed at propelling Europe to the forefront of the global AI race with a strategic move toward European technological sovereignty.
But what could truly support achieving this sovereignty? According to the panel titled “Shared Progress: How Can Open-Sourcing Speed the AI Race?”, the answer lies in Open Source AI. It’s not just about code and infrastructure, it’s about values, collaboration, and building technology that reflects cultural and democratic principles.
Here’s what we learned from that panel moderated by Niall Firth (Executive Director of MIT Technology Review), featuring Stefano Maffulli (Executive Director of the Open Source Initiative), Jean-Baptiste Kempf (CTO of Scaleway), and Laurent Solly (VP Europe at Meta).
Maffulli began by anchoring Open Source in its core philosophy: unrestricted, permissionless innovation. For AI to be considered Open Source AI, it must offer access not only to the model weights and parameters, but also to the training code, and the data (unless it’s not possible). These elements together uphold the principles of transparency and user freedom that Maffulli warns are under threat by ambiguous licensing and regional restrictions masquerading as openness. He called out practices that label models as “open” while imposing usage restrictions, especially on Europeans.
Kempf delivered a sobering but optimistic message: Europe lacks the scale of compute and capital of U.S. or Chinese giants, but it has something even more powerful—a history of building commons. “The only way Europe will not be left behind is by using Open Source,” he said, “We know how to work together.” Kempf believes the Open Source AI wave is unstoppable, likening it to the evolution of software infrastructure itself. Linux and Python, for example, began as collaborative projects by European founders and went on to become global standards. AI, he argues, is simply the next chapter.
Laurent Solly, representing Meta, emphasized their commitment to Open Source, from opening the FAIR (Fundamental AI Research) laboratory in Paris to championing tools like PyTorch and Llama. He acknowledged, however, that defining what counts as Open Source AI is a live debate, one Meta is engaged with.
Yet, Maffulli pointed out a critical flaw: Llama’s license excludes European users from using it freely, raising questions about how “open” it truly is. The debate highlights a deeper tension between Open Source as a marketing term and Open Source as a philosophy of freedom and innovation.
The panelists agreed that Open Source is not just an ideological stance; it’s an economic strategy. It allows startups, researchers, and even national governments to build powerful AI systems without relying on opaque, centralized platforms. As Maffulli warned, if Europe fails to embrace truly open practices, it risks ceding both control and innovation to firms outside of its borders .
Looking ahead, Kempf predicted a breakthrough in LLM architectures coming from the Open Source community within 18 months. He envisions a “Libre AI” era whereas models are not just open-weight, but fully transparent, reproducible, and collectively owned.
Maffulli emphasized that the biggest remaining challenge is around data provenance, data rights, and the ethical use of data. The unspoken social contract—where users give platforms their data in exchange for visibility or services—fell apart in the era of AI. Europe, with its strong regulatory frameworks and public infrastructure mindset, has a unique opportunity to lead in solving this.
The Vivatech panel made one thing clear: Open Source AI isn’t just a matter of sharing code—it’s about power, trust, and sovereignty. As France and Europe celebrate the Mistral-NVIDIA alliance and stake a claim for European leadership in AI, it must also double down on the principles that foster permissionless innovation and embrace European culture and values. Open Source isn’t just a way to catch up in the AI race. It’s how we redefine the rules of the game.
The video recording of the panel is available here:
Investing in Open Source sustainability: OSI supports OpenForum Europe’s EU Sovereign Tech Fund proposal
Introducing the new API for OSI Approved Licenses®
Receive the top news and articles of the Open Source ecosystem in your inbox, the first week of every month.
Please leave this field empty.
The content on this website, of which Opensource.org is the author, is licensed under a Creative Commons Attribution 4.0 International License.Opensource.org is not the author of any of the licenses reproduced on this site. Questions about the copyright in a license should be directed to the license steward. Read our Privacy Policy