France's €2.5 billion AI push: 81 labs, 1,000 startups, and a third phase

Les fondations posées : la phase 1 de la stratégie IA (2018-2022)

The first phase of France's national AI strategy ran from 2018 to 2022 with a budget of €1.5 billion. It was set in motion by the Villani mission, a parliamentary report led by deputy Cédric Villani that urged the government to treat AI as a national priority. That phase built the research backbone of the ecosystem: a network of interdisciplinary AI institutes (3IA) spread across the country, 180 excellence chairs in AI, funding for 300 doctoral programs, and the deployment of the Jean Zay supercomputer, now one of Europe's most powerful machines. By the end of this phase, France had 81 dedicated AI laboratories, more than any other European country, and over 1,000 AI startups had emerged from the pipeline. By contrast, the US approach has been driven largely by industry, as a coalition of 41 companies recently argued for open-weight models.

La diffusion dans l'économie : la phase 2 (2022-2025) et ses objectifs

The second phase, funded at €1 billion from 2022 to 2025, shifted focus from research to economic adoption. It worked through three levers: training and research, supporting state-of-the-art offerings, and bridging AI supply and demand. The government identified four priority domains: embedded AI (integrated into devices), trustworthy AI (reliable and transparent systems), frugal AI (energy-efficient models), and generative AI. Specific targets included training 3,500 additional students per year, supporting 400 small and midsize companies in adopting AI, and capturing 15% of the global embedded AI market. The phase aimed to turn academic strength into industrial muscle. Outside government, private actors are also funding AI policy work, as Anthropic's $40 million commitment to nonpartisan research demonstrates.

Cap sur 2030 : la phase 3 lancée en 2025

In February 2025, France launched a third phase under the France 2030 umbrella, coordinated by Guillaume Avrin, the national AI coordinator attached to the General Secretariat for Investment. The new phase centers on four priorities: strengthening computing infrastructure and critical links in the AI value chain, training and attracting AI talent, accelerating AI use cases, and building trustworthy AI. The focus on trustworthy AI aligns with the emerging global regulatory consensus. While the budget for this phase has not been detailed in isolation, it is part of the €2.5 billion total committed to AI under France 2030. The new phase acknowledges that earlier investments need to scale beyond research into deployment across the economy. Other European governments are taking similar steps, such as Greece's partnership with ElevenLabs to bring voice AI into public services.

Quels résultats ? Bilan et perspectives

The government's investment has produced these results: over 1,000 AI startups now active in France, 81 AI laboratories (the largest national network in Europe), 600 startups specialized in AI, approximately 13,500 direct jobs in AI startups, and an estimated 70,000 indirect jobs. The Jean Zay supercomputer has become a shared resource for both academic and industrial research. But the numbers also reveal gaps. The 15% global market share target for embedded AI remains ambitious, and the 400 SME adoption goal is modest relative to the size of the French economy. The third phase will test whether France can move from a research leader to an industrial competitor in AI, especially as global rivals like China's AI labs and the US accelerate their own investments.

Source: French government announcements and the national AI strategy documents.