Microsoft and Nvidia Launch Sovereign AI Framework for Enhanced Control
Microsoft has introduced a new sovereign AI framework designed to give organizations greater control over their AI workloads. The framework is built on four core principles: control, choice, flexibility, and resilience. These principles apply to data, access, governance, infrastructure, and operations. Didier Ongena, Microsoft's Public Sector & Sovereignty Leader, describes sovereign AI as the ability to design, deploy, and operate AI workloads under defined controls.
The framework offers two deployment tiers: Sovereign Public Cloud and Sovereign Private Cloud. These tiers cater to different levels of control required by various workloads. Microsoft also provides additional tools like Microsoft Foundry and Foundry Local for model development, and Azure Local for edge deployments. The company emphasizes the mantra “your data, your models, your decisions” to highlight the user-centric approach.
A joint white paper with Nvidia outlines a reference stack that includes NVIDIA AI Enterprise, Nemotron models, NVIDIA RTX PRO, and Confidential Computing. This stack is intended to protect sensitive data, models, and AI workloads during processing. Dave Salvator, Nvidia's Director of Accelerated Computing Product Marketing, notes the importance of accelerated computing and model choice for distributed AI workloads across cloud, data center, and edge environments.
Microsoft highlights Europe as a key region where sovereignty questions are particularly important. However, the company does not disclose any launch customers, contract values, or regulators involved. The framework is positioned as part of a broader strategy to build durable AI solutions that adapt to changing requirements, offering organizations the control and choice needed to use AI with confidence.