Microsoft and Intel pitch a shared stack for sovereign AI, from edge to cloud

Sovereignty moves beyond data location

According to Microsoft, AI processing is moving closer to where data is created, decisions are made and operations take place. The post gives examples: manufacturers can analyze confidential production data next to the systems that generate it, pharmaceutical companies can apply AI to clinical trials within controlled environments, and government agencies can run AI applications on sensitive information while keeping established security and operational boundaries.

Microsoft says this changes the digital sovereignty conversation. It now covers where data is stored and processed, where workloads run, who can access and operate the surrounding systems, and how those systems stay secure and operational as requirements change.

The company argues that requirements differ by workload. Data sensitivity, regulation, latency, connectivity, operational access and resilience all play a role. The result is a continuum in which organizations apply different levels of control across public cloud, customer-controlled environments and trusted partner solutions.

Azure Local, Foundry Local and Intel hardware

The post names several building blocks. Azure Local, running on Intel Xeon 6 processors, brings distributed Azure infrastructure together with a common compute foundation for customer-controlled environments. Microsoft says organizations can extend cloud and AI capabilities closer to their data while keeping consistent approaches to security, management and governance.

On the AI side, Microsoft Foundry helps organizations build and deploy AI applications and models, with options for where AI runs. Foundry Local lets organizations run AI inference in their own environments. On the same infrastructure, Intel Xeon 6 processors with Intel Advanced Matrix Extensions (Intel AMX) can accelerate AI workloads directly on the CPU, and Intel OpenVINO helps developers optimize and deploy models across supported infrastructure.

The approach also reaches beyond AI models. Microsoft 365 Local and GitHub Enterprise Local extend customer-controlled options to collaboration, source code and software development workflows, so that AI, applications, data, development and productivity workloads can sit within the same governance and operational boundary.

Verifying trust with confidential computing

A central theme is that trust for sensitive workloads needs technical protection and verifiable evidence. For supported Azure Confidential VM configurations, Intel Xeon processors provide hardware-based protections, including Intel Trust Domain Extensions (Intel TDX), which can help isolate workloads while they are in use. Hardware-backed attestation can give organizations evidence to check relevant security properties of the environment against their own requirements.

Anand Pashupathy, Vice President and General Manager of Intel Product Assurance and Security, writes in the post that on supported Azure confidential VM configurations, Intel TDX runs AI workloads in hardware-isolated trust domains, protecting models and data while in use and isolating them from the host and hypervisor. Attestation, he says, lets organizations "move from trusting their environment to verifying it."

Microsoft adds that each workload should receive protections proportionate to its risk, with isolation, encryption, attestation and operational controls applied where they add meaningful value.

What it means for organizations

The post stresses adaptability. A workload that runs in a connected environment today may face different regulatory, resilience or connectivity requirements tomorrow, and models and infrastructure will keep evolving. Microsoft says organizations should be able to change where workloads run, how they are secured and how they connect without rebuilding their management and security model each time.

For customers handling regulated, proprietary or mission-critical data, the message is that workload placement has become a practical AI decision. The post is a statement of approach and product positioning rather than a launch: it gives no pricing, availability dates or performance figures. The visible text ends as it begins a section titled "Three questions leaders should ask", which is not covered here.