OVHcloud strengthens its expertise in generative AI through the strategic acquisition of Dragon LLM

OVHcloud, a major European player in cloud computing, is strongly committed to the field of artificial intelligence by announcing the acquisition of Dragon LLM, a French company recognized for its expertise in specialized language models. This operation marks a strategic turning point for OVHcloud, which had been cautious in its investments in generative AI until now. With this acquisition, the company aims to strengthen its technological sovereignty while offering innovative services capable of addressing security challenges related to sensitive data. By combining the strengths of OVHcloud and the technical expertise of Dragon LLM, a new era opens for the deployment of generative AI applications tailored to both cloud and on-premise infrastructures of businesses.

This initiative takes place in a European context where mastering artificial intelligence technologies becomes a major geopolitical issue. OVHcloud, as a local champion, thus intends to build a sovereign alternative to international players. The acquisition of Dragon LLM perfectly illustrates this ambition, giving OVHcloud the ability to develop its own innovation lab, the “AI Lab,” dedicated to fine-tuning and training next-generation language models adapted to specific and high-added-value use cases.

A bold technological strategy: why OVHcloud chooses generative AI with Dragon LLM

OVHcloud has long adopted a measured stance towards artificial intelligence, particularly generative AI, deciding not to invest heavily in costly GPU farms. This caution is explained by the technical complexity and the still uncertain scaling requirements of the solutions. Yet, the accelerated digitalization of European companies and the advent of new data protection regulations have pushed OVHcloud to rethink its approach. The acquisition of Dragon LLM reflects this major strategic shift.

Dragon LLM, formerly known as Lingua Custodia, has distinguished itself through its expertise in generative AI applied to very specialized fields such as finance. By integrating this French gem, OVHcloud bets on a proven technology designed to be sovereign and secure. This choice is part of a desire to create a complete chain, from the design of language models to their deployment on cloud or on-premise infrastructures, guaranteeing data confidentiality.

The strength of Dragon LLM lies notably in its innovative approach, combining training of Small Language Models (SLM) up to 3.6 billion parameters and fine-tuning of larger models like Llama 3.1 or Qwen 3, which can reach up to 70 billion parameters. This advanced technicality far exceeds what most European players offer, giving OVHcloud a significant competitive advantage in a global market dominated by American and Chinese AI giants.

Moreover, the internal creation of an “AI Lab” at OVHcloud concentrates the necessary resources and expertise to innovate rapidly, test new architectures like the Mamba-Transformer, and fine-tune language models according to specific use cases. This structure will, among other things, optimize performance on applications sensitive to security and confidentiality, an essential criterion for European clients.

Dragon LLM: recognized French expertise at the heart of innovation in generative AI

Founded in 2011, Dragon LLM initially distinguished itself in specialized automatic translation in finance before turning to the booming growth of generative AI. The key shift consisted of developing its own language models adapted to sectors requiring great analytical finesse, notably banking and insurance. The company thus positioned itself as a leading player in the development of Small Language Models adapted to the needs of European markets.

Dragon LLM gained recognition by winning the prestigious Large AI Grand Challenge organized by the European Commission. This challenge required the training of a model with several billion parameters on European supercomputers, demonstrating the company’s ability to work with ultra-high-performance local infrastructures, in line with digital sovereignty requirements. This success highlights an innovative architecture, called Mamba-Transformer, recognized for its efficiency and ability to optimize the training of language models.

At the same time, Dragon LLM has fine-tuned major open-weight LLMs like Llama 3.1 and Qwen 3. The use of highly sensitive financial data in this learning process underscores the trust given to the company’s technology by prestigious banking players such as BNP Paribas, Crédit Agricole, and Natixis, who already use a platform developed by Dragon LLM called Verto. This solution goes beyond translation and allows checking the consistency of complex documents like financial prospectuses and contracts, an essential service for tendering and audits.

This increased expertise and orientation towards AI that is not only performant but also respectful of security and confidentiality standards give Dragon LLM a prominent place in the European ecosystem. Few companies today have this level of mastery, asymmetrically positioning OVHcloud in the competition for digital sovereignty in generative AI.

Secure innovative generative AI services for sensitive sectors

OVHcloud’s strategy, with the acquisition of Dragon LLM, fits into the logic of an innovative offer designed to process sensitive data securely. The services resulting from this integration will be deployable both on the group’s cloud infrastructures and on clients’ on-premise servers, a flexibility essential for companies subject to stringent regulatory constraints in Europe.

The particularity of the proposed AI solutions is based on a dual commitment: mastering the entire technological chain while guaranteeing data sovereignty. This positioning opens the door to critical uses in finance, health, defense, or energy, where information protection is non-negotiable. For example, banks will be able to rely on generative models capable of analyzing large and complex contracts while respecting the confidentiality of internal data.

The challenge is all the more crucial as European legislation, through texts such as GDPR, imposes strict standards on the processing of personal and sensitive data. OVHcloud and Dragon LLM therefore combine their strengths to implement solutions compliant with these requirements, with the possibility of deploying hybrid environments, combining public cloud and private infrastructures, to meet specific needs.

These offerings also target organizations that wish to maintain direct control over their environments. For example, a company in the energy sector may prefer to host its critical data locally while benefiting from generative AI capabilities to optimize industrial processes, detect anomalies, or automatically generate detailed reports.

• Adaptability: deployment on public, private, or hybrid cloud according to needs.
• Security: advanced encryption and GDPR compliance ensured.
• Sovereignty: use of European technologies and local hosting.
• Performance: models refined for specific and large-scale sectors.
• Support: integrated technical support via OVHcloud.

Impact of the acquisition on current Dragon LLM clients: a major change in perspective

For Dragon LLM’s historical clients, the takeover by OVHcloud represents a significant evolution. On the one hand, consolidation with a major European player now guarantees better economic and technical stability. This assurance is vital for companies that had so far depended on the viability of an independent startup in a very competitive sector.

On the other hand, the gradual integration of Dragon LLM services into the OVHcloud ecosystem could lead to changes in the way solutions are deployed and maintained. Migration to OVHcloud’s cloud platforms is conceivable, with benefits in integration, scalability, and simplification of IT architectures.

Nevertheless, this transition also requires an adaptation phase for clients, who may need to revisit certain contractual, technical, or operational aspects. For example, switching to a new billing model or to different support and maintenance systems may cause organizational adjustments. This reality illustrates a well-known issue in any acquisition: the need to support clients to minimize any negative impact.

The risk of a strategic product reorientation should not be underestimated. OVHcloud could favor certain innovations or focus its developments on particular segments, which could accelerate some features or slow down others, causing frustrations among users accustomed to a different pace or offering.

Aspect	Before acquisition (Dragon LLM)	After acquisition (OVHcloud)
Financial stability	Startup dependence, higher risks	Strong group support, better sustainability
Technological integration	Third-party infrastructures, independence	Connection to OVHcloud ecosystem, possible migration
Customer support	Dedicated startup support, limited	Strengthened support by OVHcloud
Product roadmap	Flexible, startup-oriented	Prioritization by OVHcloud, possible reorientation
Contractual conditions	Own, personalized conditions	Possible evolutions and standardization

A strategic shift for OVHcloud in the race for AI innovation

This acquisition of Dragon LLM reflects both a change of stance and an acceleration in OVHcloud’s artificial intelligence strategy. Internally, the launch of the “AI Lab” marks a desire to make applied research and innovation a priority investment. This structure will not only refine existing models but also design new architectures adapted to the specifics of the European market and its sovereignty requirements.

It is also a strong signal sent to the market, demonstrating OVHcloud’s ambition to compete with global leaders. This approach takes place in a context where American giants such as Google, Microsoft, or OpenAI largely dominate the sector, imposing their technology and standards. By strengthening its expertise thanks to Dragon LLM, OVHcloud seeks to offer a competitive alternative respectful of European values, particularly in data protection and technological ethics.

This orientation could also indirectly influence public policies by strengthening the technical capacities available on European soil. The current geopolitical context advocates greater digital autonomy, where mastering technologies like artificial intelligence becomes a strategic priority. In this framework, close collaboration with public players and research institutions will amplify the innovation lab’s impact.

Future prospects: towards a sovereign cloud enhanced by generative AI

Beyond the immediate integration of Dragon LLM’s technologies, OVHcloud paves the way for a new generation of cloud services enriched by generative AI. The goal is to offer its clients, whether companies, administrations, or institutions, integrated solutions combining the power of cloud computing with the sophistication of specialized artificial intelligence.

Potential projects include creating customized platforms for automated document analysis, generating adapted summaries, or intelligent virtual agents capable of assisting users in their daily tasks. These innovations will emphasize security, data traceability, and compliance with European regulations.

This vision fits into the dynamics of a market where demand for advanced AI capabilities continues to grow. Economic actors seek solutions capable not only of improving their productivity but also of strengthening customers’ and partners’ trust in their digital infrastructure. OVHcloud, with Dragon LLM, thus positions itself as a future leader for sovereign AI in Europe.

• Seamless integration between cloud and generative AI.
• Priority given to security and data confidentiality.
• Continuous development of models adapted to regulated sectors.
• Tailor-made offers to meet specific needs of European clients.
• Strengthening collaborations with research centers and institutions.

Ethical issues and technological sovereignty at the heart of OVHcloud’s strategy

The rapid rise of generative AI raises many ethical questions, to which OVHcloud intends to respond proactively. Data security, algorithm transparency, and respect for privacy are fundamental axes in the development of integrated solutions. The desire to build a European sovereign AI goes beyond mere technical implementation to touch a major societal and political commitment.

For OVHcloud and Dragon LLM, responsibility goes beyond performance and results: it involves ensuring that generative AI systems do not discriminate, do not bias decisions, and comply with a strict legal framework. Particular attention is paid to the traceability of data used in training as well as the possibility to audit algorithms to guarantee full compliance with European standards.

This approach also takes into account risks related to technological dependence on foreign powers. By strengthening its local expertise, OVHcloud consolidates Europe’s digital resilience and offers a genuine alternative to dominant offerings, often criticized for their opacity and control of user data.

In this context, future collaborations with standardization bodies, universities, and research laboratories will be crucial in developing robust ethical frameworks, which could serve as references at the European level. A collective awareness thus accompanies this major technological advancement.

Towards a new stage in European cloud computing with OVHcloud and Dragon LLM

With this ambitious acquisition, OVHcloud takes a major step in consolidating its industrial offering around artificial intelligence. The introduction of Dragon LLM’s skills raises innovation capabilities to a level that meets the demands of a rapidly changing European market. This movement also opens up prospects for strategic partnerships and joint developments likely to enrich the local digital ecosystem.

Positioning itself as a provider of sovereign cloud solutions integrating advanced generative AI capabilities gives OVHcloud a particularly strategic competitive advantage. European institutions and companies are more than ever looking for a partner capable of combining technological performance with compliance with regulatory constraints, a balance difficult to achieve globally.

The coming together of two pioneering players in their respective fields illustrates a dynamic of cooperation and scaling-up essential to maintaining European competitiveness in the coming years. OVHcloud and Dragon LLM thus invite us to imagine a future where technology goes beyond mere innovation to embody a sustainable, secure, and ethical approach.