

TLDR
Key Metrics:
NTT DATA and Haystack by deepset are launching a new collaboration centered on a Sovereign AI Transformation model: an approach for moving AI from isolated pilots to governed, production systems in regulated and complex environments.
The model addresses a pattern that has become common across the public sector and regulated industries. Organizations have run AI pilots, often successfully, but struggle to scale them into production while meeting their requirements for data sovereignty, governance, and regulatory compliance. The result is fragmentation: many pilots, few production systems, and limited institutional learning.
The Sovereign AI Transformation model is designed to close that gap by combining a sovereign AI platform with a structured operating model and institutional governance.
First deployment: AI@EC
The first deepset deployment is for the AI@EC Platform: The European Commission's sovereign AI foundational pillar for moving from isolated pilots to governed, production systems across the institution.
AI@EC Platform is well supported by DIGIT Centre of Excellence in Data Analytics & AI, which coordinates AI demand across the Commission, sets best practices, and advises stakeholders on effective AI adoption.
How the model works
The partnership brings together three layers:
- Haystack by deepset provides the sovereign AI platform: self-hosted, modular, with full control over data, models, and deployment. Built on the open-source Haystack framework, the platform supports the development and operation of AI applications including agents, retrieval-augmented generation, search, and assistants.
- NTT DATA brings the operating model: structured processes, implementation support, and the path from use case to production. NTT DATA's delivery capability operationalizes the model and provides the institutional knowledge needed to scale AI adoption sustainably.
- The institution provides governance and coordination: managing demand, enforcing standards, and supporting teams across the organization. In the case of AI@EC, this role is held by DG DIGIT's Data & AI Center of Excellence.
Together, these three layers create a repeatable system for building AI applications under the compliance and sovereignty requirements that regulated organizations operate under.
Designed to be replicable
While the first deployment is with the European Commission, the model is designed to be replicable. Any public sector or regulated organization facing the same challenges — fragmented pilots, governance requirements, and sovereignty constraints — can adopt the same approach.
The model is platform-agnostic in its delivery shape: the operating model and governance layer adapt to the institution's existing structure, while the sovereign AI platform provides a consistent technical foundation.
About the partnership
NTT DATA and Haystack by deepset are committed to helping European institutions build and scale AI on their own terms. The collaboration reflects a shared view that sovereignty and scale are not trade-offs, and that the right combination of platform, delivery, and governance can deliver both.
To learn more about the Sovereign AI Transformation model, or to discuss how it might apply to your organization, contact us.
Get started with Customized AI.
Table of Contents
Explore Related Content
.png)
September 9, 2026
deepset Named In Sifted 100 Startups in DACH & CEE
The Sifted 100: DACH & CEE 2026 recognizes the fastest-growing startups and scaleups across the region. deepset is proud to be included in this year’s ranking, reflecting our continued growth as we help organizations build and scale sovereign AI systems, agents, and applications.
news

September 3, 2026
Materna and deepset Strengthen Sovereign AI Made In Germany
Materna and deepset are combining their expertise to advance the sovereign use of artificial intelligence in the public sector and regulated industries.
news

August 4, 2026
Private AI Cloud: Innovation in the Compliance Zone
How can financial service providers use a private AI cloud to move from pure AI consumer to producer of their own intelligence? That move is a strategic decision, caught between dependence on the large AI providers and self-determination. But no single vendor delivers the full stack alone, which is why HPE, NVIDIA, deepset, QuantPi and secunet are pooling their capabilities into a shared platform rather than competing as individual vendors.
news






