A European collaboration for trustworthy AI in critical industries
On May 13, 2024, Capgemini, SAP, and Mistral AI announced a strategic partnership aimed at developing generative artificial intelligence solutions tailored to regulated industries. This joint initiative aims to provide high-performance, sovereign AI services that meet the requirements of critical environments such as healthcare, finance, defense, and public administration.
A technological solution to the business challenges faced by regulated industries
Regulated industries have specific requirements regarding confidentiality, traceability, robustness, and compliance. This joint initiative enables the integration of advanced language models (Mistral 7B, Mixtral) into SAP infrastructures through Capgemini’s expertise. These systems are fully customizable and can be deployed on-premises or in a private cloud.
The proposed solutions are designed for specific use cases:
- Finance: Automated generation of regulatory reports, AI-enhanced internal audit.
- Healthcare: Automated generation of anonymized medical summaries for healthcare providers.
- Public sector: Improving user experience through multilingual, transparent, and audited chatbots.
A sovereign AI designed for Europe
Unlike most LLM offerings hosted by GAFAM companies, this partnership claims full digital sovereignty. The models are open source, the infrastructure can rely on European-certified clouds (Gaia-X, NumSpot, Bleu), and data localization is guaranteed.
This strategic direction is part of Europe’s commitment to making its technological infrastructure self-sufficient. The initiative is therefore fully aligned with the requirements of the GDPR and anticipates the provisions of the AI Act1.
Ethical governance embedded from the outset
The ethical approach adopted is not defensive but constructive. The three partners emphasize proactive governance, integrated into every stage of development:
- Full traceability of requests and responses.
- Open auditability thanks to the use of open-source code (Mistral).
- Continuous human oversight to ensure that the algorithm does not deviate from its intended path.
- Joint ethics committees comprising engineers, lawyers, and subject-matter experts.
This approach, which aligns with the principles of "trustworthy AI," seeks to build trust through transparency and inclusivity, rather than simply reacting to risks.
Adoption made easier by interoperability
One of the key benefits of the partnership is the seamless integration of solutions into existing infrastructures. The architecture is based on standards (ONNX, Docker, SAP BTP API) that enable strong interoperability with ERP systems, document repositories, and business tools.
Adaptability is key for CIOs who want to avoid vendor lock-in while complying with internal security and sovereignty policies.
Rapid adoption is expected in critical infrastructure
The first pilot deployments are scheduled to begin in the second half of 2025, particularly in European government ministries, central banks, and hospitals. Capgemini estimates that 25% of large European companies operating in regulated environments will adopt sovereign AI by the end of 20262.
This momentum is being driven by the growing volume of public and private investment in European AI infrastructure. The Gaia-X initiative, sovereign data centers, and platforms such as NumSpot and Bleu could host these solutions while ensuring a local presence.
Toward a new European standard for business AI?
This partnership could well set the stage for a European industrial model for AI adoption, combining innovation, compliance, and performance. Capgemini estimates that by 2026, one in four companies operating in regulated environments will adopt sovereign AI models.
This initiative, supported by public policy, could foster the emergence of an ecosystem of responsible, interoperable AI solutions tailored to business needs. This momentum could inspire other technological collaborations across the continent.
References
1. SAP Newsroom. (2024). Mistral AI, SAP, and Capgemini to collaborate on sovereign AI for enterprises.
www.news.sap.com
2. Capgemini. (2024). Capgemini, SAP, and Mistral AI join forces to develop AI solutions for regulated industries.
. www.capgemini.com
