Microsoft has set out the next stage of its Sovereign Cloud offering in a post on the Azure Blog. Most of it covers access controls and resilience guidance for regulated customers. The parts that involve AI are about running models and AI-assisted tools inside environments the customer operates.
What is changing for AI workloads
According to Microsoft, Foundry Local will become generally available at Ignite for connected and disconnected operations on Azure Local. The model catalog is expanding to include Microsoft, partner, open-weight and open-source models.
Microsoft also describes local inference and agentic retrieval. Together they let users ask questions about local enterprise data in natural language and get answers grounded in that data. An integration with Microsoft 365 Local extends this to locally operated SharePoint and Exchange servers. It is in preview for approved customers only.
For developers, GitHub Enterprise Local becomes generally available at Ignite as a validated GitHub Enterprise Server appliance on Azure Local. Microsoft says GitHub Copilot can use locally hosted models for inference, so AI-assisted coding stays close to the code and data.
Protection and isolation
Confidential virtual machines on Azure Local are already generally available. They use hardware-encrypted memory and isolation to protect data while it is processed, and workload attestation adds a check before sensitive data or keys are released. Azure Local will also support multitenancy, so one set of infrastructure can serve several tenants with separate resources, governance and management.
The wider announcement
Microsoft frames the package as a continuum of control across Sovereign Public Cloud, Sovereign Private Cloud and a partner ecosystem. Public-cloud additions include:
- a Sovereign Control Panel entering public preview;
- Data Guardian and Production Touch Ledger, rolling out in phases for EU and EFTA customers;
- external key management with Azure Managed HSM, now generally available;
- Azure Confidential Clean Rooms, in limited preview.
On the private side, the ability to move workloads from connected to disconnected operation enters public preview. DocumentDB and Blob Storage on Azure Local are also entering public preview.
What the source does not say
The post gives no pricing, no list of specific models in the catalog and no regional availability for the AI features. Those details are not confirmed here. It is a vendor announcement, and the capabilities have not been independently tested.
Who should pay attention
The audience is organisations that must keep models, data and processing under their own control. Everyone else can treat it as a signal. Local, disconnected AI over internal data is now a supported deployment option for a large vendor, not only a custom project.