France and Germany have announced a new strategic partnership with Mistral AI and SAP SE to build what they frame as a sovereign AI infrastructure for public administration.
As shown in the chart below, the US and China account for roughly two-thirds of newly funded AI start-ups worldwide. This joint initiative, revealed on 18 November, aims to reduce Europe’s technological dependence on non-European providers. The announcement took place at the Summit on European Digital Sovereignty in Berlin, where France and Germany emphasised strengthening European innovation and competitiveness.
An analysis at Politico highlights that Macron and Merz voiced strong hopes to remove bureaucratic hurdles in the EU’s digital rulebook.
“If we let the Americans and the Chinese have all of the champions, one thing is certain: we may have the best regulation in the world, but we won’t be regulating anything,” said Macron.
By mid-2026, the commitment will be finalised through the signing of a binding Framework Agreement. The rollout of initial use cases will begin between 2026 and 2030. These dates align strategically with the next Multi-annual Financial Framework (MFF) currently being negotiated between the Council of the European Union and the European Parliament.
The key pillars of the joint initiative
According to the press release, the partnership focuses on four pillars designed to reshape public administration:
- AI-native sovereign Enterprise Resource Planning (ERP) systems tailored for French and German administrations
- Automated financial management tools using advanced AI
- AI agents for civil servants and citizens, including scenario simulators, compliance tools, and guidance chatbots
- Joint innovation labs and workforce upskilling programs
A promising step towards European sovereign AI
The Draghi report highlights that Europe faces a competitiveness crisis. According to his analysis, 61% of total global funding for AI start-ups goes to US companies, 17% to those in China and just 6% to those in the EU. Such low level of funding in European AI start-ups limits growth, thus affecting Europe’s digital sovereignty.
A report published last April by CIGREF (the IT Club of France’s Largest Enterprises) estimates that the cost of digital dependence is at around €265 billion per year, reflecting a massive transfer of wealth to mostly the U.S. In the current scenario, Europe risks becoming a permanent importer of critical digital infrastructures.
Since 2019, several European leaders have urged the Commission to move faster on building the Union’s digital sovereignty. The World Economic Forum notes that this drive is closely linked to the EU’s emphasis on protecting individual rights, which underpins its push for greater autonomy in the digital sphere.
The Franco-German initiative could help bridge the European AI gap while allowing the Union to maintain its desired standards. Meanwhile, Draghi also warns about the risk of fragmentation. If this policy remains restricted to two countries, the scaling up of Mistral AI and SAP SE could be hindered. Smaller countries may lack the resources to build parallel sovereign AI infrastructures, creating a “core versus periphery” dynamic. Such asymmetry could undermine the Single Market, complicate interoperability, and weaken Europe’s bargaining power with U.S. and Chinese cloud providers.
Critiques, challenges and risks
The Draghi report also highlights high energy costs as an obstacle to the AI sector’s growth. As shown in the graph below, electricity prices are higher in Europe than in the US or China. Around half of European companies see this as a major barrier to investment – 30% more than US firms.
Training and running AI models is highly energy-intensive. Draghi posits that by 2030 the electricity consumption of data centres in the EU is likely to rise by 28% compared to 2024 levels. As such, the EU must find ways to lower energy prices to succeed in AI.
Beyond macroeconomic constraints, there are also operational and legal risks linked to the way AI is deployed in public-sector work. As an example, the Catalan news outlet Diari Ara disclosed that Deloitte returned thousands of dollars to the Australian government over a report with AI errors. This incident, which occurred in October 2025, illustrates that the use of poorly supervised AI tools may produce mistakes with financial, legal, and reputational damage.
Therefore, the initiative needs to consider the risks AI may pose to public administration, particularly with regard to the implementation in financial management and guidance chatbots. Therefore, Europe should balance its competitiveness strategy with regulation of technological tools.
Conclusion
Ultimately, the Franco-German sovereign AI initiative captures the central tension of Europe’s digital moment: the continent must move fast, but together. If expanded beyond two capitals, the project could become a genuine European backbone for public-sector AI. If not, it may merely underline the fragmentation Draghi warns against. As Europe enters a decade defined by technological power, the real test will be whether this partnership becomes a continental catalyst or just another bilateral experiment.
You may also enjoy reading…
Europe’s chip race: inside the Dutch Nexperia takeover
Geopolitical risk analysis in an uncertain world: the case of CogniTrends
Technology, Geopolitics, and Business: Professor Andrea Colli’s Insights