G42-Backed Domyn Launches AI Model for Regulated Industries
Domyn introduced Domyn Large, a 260B parameter reasoning model, through Microsoft Foundry on April 15. At the same time, G42 supports the company, reinforcing its expansion strategy.
Moreover, Domyn is raising new capital while targeting banks, pharmaceutical firms, governments, and defence operators. These sectors require strict data control, so the company emphasizes secure and auditable deployments. As a result, Domyn positions itself as a provider of controlled AI infrastructure rather than open-access models.
Uljan Sharka outlined a $10 billion, three-year investment plan for an Italian supercomputer project with NVIDIA and G42. Therefore, the company links its growth to large-scale compute investments.
Ownership Model and Enterprise Use Cases
Domyn frames ownership as its core commercial advantage. Instead of offering rented access, it allows clients to keep models, training data, and infrastructure within governed environments. Consequently, organizations gain full control over sensitive operations.
In addition, Domyn Large supports mission-critical enterprise tasks with reasoning controls, tool integration, and compliance with the EU AI Act. It also supports more than 50 languages, which broadens its applicability. Furthermore, clients can adapt models using proprietary data through continual pretraining.
The company targets industries with complex data needs. For example, banks can analyze transactions and detect anomalies, while insurance firms can process claims and risk models. Similarly, utilities and manufacturers can evaluate maintenance records and production workflows.
Meanwhile, Microsoft positions Foundry as a unified system for building and deploying AI agents at scale. Therefore, Domyn gains access to enterprise customers seeking governance, secure deployment, and operational control.
Expanding Compute Ambitions and Industry Reach
At the same time, Domyn’s strategy relies heavily on large-scale compute infrastructure. In May 2025, G42 revealed an Italy-based compute cluster using thousands of NVIDIA Blackwell GPUs. Initially, the project involved a $1 billion phase, yet current plans now target $10 billion, signaling a major scale increase.
Additionally, NVIDIA identified Domyn as a collaborator on an AI factory powered by 5,760 Grace Blackwell GPUs. Domyn also describes Colosseum as infrastructure designed for models exceeding one trillion parameters. As a result, compute capacity will directly influence model performance and deployment scale.
Healthcare remains a key focus area. Sharka explained that pharmaceutical companies can run parallel experiments and shorten research timelines using the platform.
However, Sharka also noted that future AI architectures may require less concentrated compute. Nevertheless, Domyn continues to prioritize building its own supercomputing capacity. Therefore, funding and infrastructure access remain central to its long-term growth strategy.
