Materna and deepset Strengthen Sovereign AI Made In Germany

(Translated from the Original Article)
‍
Dortmund/Berlin, [Wednesday 2nd September] – Materna and deepset are combining their expertise to advance the sovereign use of artificial intelligence in the public sector and regulated industries. The joint offering combines Materna’s experience in the digital transformation of complex administrative and IT landscapes with the technology of deepset, a young, innovative AI company from Berlin.

‍

deepset stands for European AI development with strong technological relevance. Founded in 2018, the company established Haystack, an open framework that has been designated as the AI orchestration standard in the Germany Stack. Haystack provides the technological foundation for connecting language models with an organization’s own knowledge in a controlled manner and using them to develop sovereign AI applications, without transferring sensitive information to external model environments in an uncontrolled way. This gives deepset a key capability for building trustworthy AI applications flexibly, transparently, and independently of individual vendors. Haystack Enterprise Platform builds on this value with the capabilities regulated organizations need to scale AI, including development workflows, operational controls, and governance tools that combine open-source flexibility with enterprise-grade control.

‍

This is particularly relevant for government agencies and organizations with stringent requirements for data protection, traceability, and control. They can use AI without giving up sovereignty over their data, knowledge sources, and operating environments. Here, sovereign AI means that sensitive information remains under control, results are based on traceable sources, and applications can be embedded within existing governance, security, and compliance structures. Applications can be implemented across different deployment models, from sovereign cloud environments to operation within an organization’s own or otherwise controlled infrastructure.

‍

At the heart of the partnership is the journey from an idea to a reliably usable AI solution. Many organizations are currently testing individual AI applications but face the challenge of integrating them securely, economically, and sustainably into day-to-day work. Materna and deepset are addressing this with an approach that combines technological openness and operational control with proven implementation expertise.

‍

Materna brings concrete experience from the production deployment of sovereign AI solutions at federal and state government agencies, for example, AI-based agent and assistant systems that support professionals with knowledge-intensive tasks, securely make internal information accessible, and operate within controlled IT environments. Haystack is also already being used in production by municipalities and federal ministries, for example to improve application processes.  It is also the engine of the co-developed AI platform, AI@EC, within the European Commission.

‍

“Materna brings many years of experience in consulting, implementation, and operating solutions for the public sector,” says Johannes Rosenboom, SVP Sales Public and KRITIS Sector at Materna. deepset complements this strength with modern AI technology, rapid development capabilities, and a clear focus on production-ready applications. Together, the partners are creating a foundation that enables public administrations and regulated organizations not merely to experiment with AI, but to deploy it responsibly and at scale.

‍

“Sovereign AI only becomes meaningful when organizations can put it into production on their own terms. Together with Materna, we can bring open, European AI technology into the environments where control, trust and reliability matter most,” says Milos Rusic, CEO and Co-founder, deepset

‍

The partnership underscores the ambition to bring digital sovereignty and innovation together: with European technology, a strong integration partner, and clear value for organizations seeking to introduce artificial intelligence into their processes securely and effectively.

‍