Microsoft and Mistral have announced a significant expansion of their strategic alliance to help companies and regulated sectors adopt next-generation AI with more options, greater control and operational consistency.
Both companies are integrating Mistral’s most advanced and efficient models on the Microsoft platform, including Microsoft Foundry, Copilot Studio and Azure, so that customers can build and run AI solutions in a wide range of operating environments: from scaled cloud deployments to completely disconnected, customer-controlled operations.
In Europe and other regulated markets, organizations are demanding access to next-generation AI without losing control over their data, operations and critical workloads. This alliance reinforces the Microsoft Sovereign Cloud approach by combining Mistral’s most advanced models with security, compliance and Microsoft’s platform, which spans from the cloud to the edge, giving customers more options on how and where they deploy their AI solutions.
“Europe must be able to access the world’s most advanced AI without giving up control over its data, its operations or its digital future,” said Brad Smith, vice president and president of Microsoft. “By incorporating Mistral’s pioneering European models into our sovereign cloud portfolio and enabling them in public cloud, cloud-connected and fully disconnected environments, we are delivering on our European Digital Commitments and offering customers a trusted foundation to develop and manage AI according to their own needs.”
“Our mission has always been to put the most advanced AI within the reach of any organization, while maintaining control over the technology,” says Arthur Mensch, co-founder and CEO of Mistral. “With Microsoft as a partner, our models reach companies and public institutions on a global scale, through a trusted platform for the most demanding regulated workloads and with availability wherever our customers operate.”
The future of AI in Europe: increased GPU capacity
The agreement builds on a new multi-million dollar commitment aimed at expanding AI infrastructure in Europe. Mistral will bring new GPU capacity by deploying thousands of next-generation NVIDIA Vera Rubin GPUs, with the goal of expanding the availability of AI computing for customers and offering a shared platform for training, inference and large-scale deployment.
The agreement strengthens the AI infrastructure in Europe and helps Microsoft respond to the growing demand for cloud and artificial intelligence services. In line with Microsoft’s flexible approach to its global infrastructure – combining owned data centres, leased facilities and strategic collaborations with third parties – it expands Microsoft’s capacity in Europe and supports the European Digital Commitments announced in 2025.
“Agentic AI is driving unprecedented demand for high-performance, energy-efficient AI infrastructures,” said Ian Buck, vice president of Hyperscale and High Performance Computing at NVIDIA. “With the large-scale deployment of NVIDIA Vera Rubin systems, Mistral and Microsoft will provide customers with the computing foundation they need to create and run the next generation of AI in Europe and the rest of the world.”
Next generation AI -AI Frontier-
At the platform layer, the latest Mistral models, Medium 3.5 and OCR 4, are now available in Microsoft Foundry, giving developers access to next-generation models in a unified environment to build, customize, and deploy AI applications. Mistral Medium 3.5, an open weights model, is built into an Azure managed environment, enabling developers and enterprises to build, customize, and deploy AI applications with greater control, sovereign deployment options, and predictable, cost-efficient scalability. OCR 4 supports structured document processing flows and agentic workflows, and both models can be applied in agentic applications, automation and industry-specific solutions, using tools and workflows already established in the Foundry platform.
At the application layer, both companies have incorporated the Mistral Medium 3.5 model into Copilot Studio, combining the flexibility of the model with enterprise governance capabilities. This allows teams to choose the most appropriate model for each scenario, while maintaining control over how and where data is processed.
Organizations can develop AI applications using the same models, tools, APIs, and workflows in Microsoft Foundry and Foundry Local. This gives teams a unified way to build, customize, and operate AI applications, regardless of the environment in which they ultimately run.
Microsoft Foundry offers a development platform to discover, build, and deploy models and agents in the cloud. Foundry Local brings that development and execution experience to Azure Local, enabling organizations to bring AI closer to their data, users, and operating environments. Together, both solutions help reduce the need to redesign applications for each deployment scenario, while offering customers more flexibility to meet their sovereignty, latency and resiliency requirements.
Flexible deployment with a unified operating model: Azure and Azure Local
Organizations increasingly need different levels of control based on the sensitivity of their workloads, regulatory obligations, and operational requirements. Azure and Azure Local offer a unified platform that allows you to deploy AI in a wide range of operating environments:
- Cloud: Azure-hosted deployments to take advantage of cloud scale, agility, and access to the platform’s latest innovations.
- Connected to the cloud: On-premises, customer-controlled Azure environments that remain connected to Azure services and operations when needed.
- Completely offline: Azure On-Premises deployments that can operate independently, without external connectivity, in highly sensitive, restricted, or critical environments.
In all these operating models, clients can use Mistral models on a common platform, with a unified management experience. This helps regulated organizations avoid a fragmented AI architecture, while allowing them to maintain the level of control, resilience and connectivity their workloads require.
At the application layer, both companies have incorporated the Mistral Medium 3.5 model to Copilot Studio
For regulated sectors that require strategic autonomy, this offers concrete advantages. Such customers can apply AI to sensitive workflows, aligning data, operations and access controls with their specific requirements. Critical infrastructure providers can maintain AI capabilities where resiliency and continuity of service are essential. Industrial and manufacturing organizations can analyze production, quality, and operations data locally when latency, intellectual property protection, export controls, cybersecurity, and supply chain resilience dictate deployment requirements. Healthcare organizations can support AI-enabled workflows in contexts where privacy, data residency, clinical continuity, and regulated data processing are critical requirements.
What it brings to customers
As part of the expanded collaboration agreement, both companies are working on a joint go-to-market plan and will collaborate on various business opportunities in Europe and globally. Mistral and Microsoft are also strengthening their alliance to accelerate customer adoption, by financing proofs of concept, granting Azure credits and organizing workshops aimed at driving AI innovation with customers.
Organizations in financial services, industry, healthcare and other regulated sectors are already using AI in environments where control and resilience are imperative. Thanks to this alliance, they will be able to create AI applications in Microsoft Foundry and run them in Azure or Azure Local, using Mistral models in cloud, cloud-connected or completely disconnected operating environments.
Microsoft and Mistral will continue to work to respond to customer needs, advancing the models, development experience, development platform and European AI infrastructure that make all this possible.
