- OVHcloud has signed a binding agreement to acquire Dragon LLM, a French AI company that builds specialized language models for regulated industries.
- The acquisition marks OVHcloud’s first ever and launches its new AI lab focused on training and fine-tuning sovereign LLM models.
- Dragon LLM, originally founded as Lingua Custodia for the financial sector, won the European Commission’s Large AI Grand Challenge.
- The combined offering will let enterprises deploy generative AI for sensitive data on-premises or in OVHcloud’s European cloud infrastructure.
What Happened
OVHcloud, Europe’s largest cloud infrastructure provider, announced on March 25, 2026, that it has signed a binding agreement to acquire Dragon LLM, a Paris-based developer of specialized generative AI models. Financial terms of the deal were not disclosed.
Dragon LLM was originally founded under the name Lingua Custodia with a focus on AI applications for the financial sector. The company later expanded into a broader fine-tuning platform designed for regulated industries, developing what it describes as high-performance, cost-effective AI models that can be deployed on local infrastructure.
The acquisition is OVHcloud’s first in the company’s history. Alongside the deal, OVHcloud is launching a dedicated AI lab that will use Dragon LLM’s technology as its foundation for building new generative AI services aimed at enterprise customers handling sensitive data.
Why It Matters
The deal reflects a growing push among European technology companies to build AI capabilities that keep data within European jurisdictions. Enterprises in regulated sectors like finance, healthcare, and government face strict data residency and sovereignty requirements that can complicate the use of AI services hosted by US-based hyperscalers.
By combining OVHcloud’s data center infrastructure across Europe with Dragon LLM’s model fine-tuning expertise, the merged entity aims to offer a vertically integrated sovereign AI stack. Customers would be able to train and deploy custom language models without their data leaving European-controlled infrastructure.
The timing aligns with increasing regulatory pressure in the EU around AI governance. The EU AI Act, which entered into force in 2024, creates compliance incentives for organizations to use AI systems where they can audit the full data pipeline, from training data to inference, within a single jurisdictional framework.
Technical Details
Dragon LLM specializes in fine-tuning large language models for specific business domains rather than training foundation models from scratch. The company’s platform is designed to produce smaller, domain-specific models that can run on local infrastructure, reducing both cost and latency compared to general-purpose cloud-hosted models. This approach prioritizes energy efficiency and deployment flexibility over raw model scale, a tradeoff that aligns with European regulatory preferences for proportionate and auditable AI systems.
The company’s approach earned recognition from the European Commission, which selected Dragon LLM as a winner of its Large AI Grand Challenge. That program was established to identify and support European AI companies building competitive alternatives to US and Chinese AI systems.
OVHcloud’s new AI lab will focus on training and fine-tuning sovereign LLM models using Dragon LLM’s technology stack. The resulting services will be deployable both in OVHcloud’s cloud and on customer premises, giving organizations flexibility in how they manage sensitive data. This hybrid deployment model is particularly important for financial institutions and government agencies that may require on-premises inference for certain workloads while using cloud resources for model training.
Who’s Affected
The acquisition is most relevant to European enterprises in regulated industries, particularly financial services, where data sovereignty requirements are strictest. Organizations currently using or evaluating AI solutions that require European data residency gain a potential new option.
For OVHcloud’s existing cloud customers, the integration signals that AI-native services will become a more prominent part of the platform. Competitors in the European sovereign cloud space, including Scaleway and Deutsche Telekom’s T-Systems, face a more capable rival. US hyperscalers operating European regions, including AWS, Azure, and Google Cloud, may also see increased scrutiny from customers weighing whether regional data center presence is sufficient for sovereignty requirements or whether European-owned infrastructure is preferred.
What’s Next
The deal remains subject to closing conditions, and OVHcloud has not announced an expected completion date or regulatory approvals required. The company has also not disclosed how Dragon LLM’s services will be priced, what model sizes the AI lab will target, or when the first sovereign AI products will reach general availability. Details are available in OVHcloud’s official announcement.
Related Reading
- Kaltura Acquires PathFactory for $22M to Build Agentic Video Experience Platform
- Ente Launches Ensu, a Free Open-Source On-Device LLM App
- Lightfeed Open-Sources TypeScript Library for LLM-Driven Web Extraction
- Bluesky’s Attie Lets Users Build Custom Feeds With Plain-Language Prompts
- Meta Acquires Moltbook, a Social Network Built Exclusively for AI Agents
