Microsoft scales Azure Local to thousands
- Microsoft said on April 27 Azure Local can now run sovereign private cloud deployments spanning thousands of servers inside one customer-controlled boundary. - The update arrives in Azure Local version 2604, adding disaggregated compute-and-storage designs, Fibre Channel SAN support, and GPU-ready infrastructure for local AI workloads. - Microsoft is recasting sovereignty as deployable on-prem cloud operations, not a separate public cloud region. (blogs.microsoft.com)
Microsoft said April 27 that Azure Local can now scale to thousands of servers inside a single sovereign environment. (blogs.microsoft.com) Azure Local is Microsoft’s on-premises platform for running Azure-managed infrastructure on hardware a customer owns. In the new pitch, that hardware can span large datacenters, industrial sites, and edge locations while staying inside one administrative boundary. (azure.microsoft.com) (blogs.microsoft.com) The product update is Azure Local version 2604, released April 27 as generally available for the new larger-scale disaggregated model. Microsoft said it lets customers separate compute from storage instead of buying one tightly coupled cluster. (techcommunity.microsoft.com) In plain terms, Microsoft is moving Azure Local from a small-box branch-office design toward a multi-rack datacenter design. The company said the new architecture can grow beyond the old 16-node cluster ceiling and up to thousands of nodes in sovereign, government, defense, and regulated environments. (techcommunity.microsoft.com) (techzine.eu) The storage piece matters because customers can now attach external storage area network, or SAN, systems over Fibre Channel instead of relying only on local disks. Microsoft said that support lets organizations reuse existing storage investments and scale compute and storage independently. (techcommunity.microsoft.com) (learn.microsoft.com) Microsoft is also tying the update to artificial intelligence workloads that cannot leave a country, agency, or facility. The company said larger fault domains, infrastructure pools, and GPU support let customers keep sensitive models and operational data on customer-controlled systems. (blogs.microsoft.com) That AI framing has been building for weeks. In March, Microsoft said Azure Local would support sovereign AI systems on NVIDIA platforms, including current RTX PRO and future Rubin-based systems, for governments and regulated industries. (techcommunity.microsoft.com) The company is also widening the hardware ecosystem around the product. Microsoft said Azure Local is available with validated compute and enterprise storage platforms from partners including DataON, Dell Technologies, Hitachi Vantara, HPE, Lenovo, and NetApp. (blogs.microsoft.com) Not every Azure service comes along for the ride. Techzine noted that services such as Microsoft Fabric, Azure Cosmos DB, and Azure OpenAI Service are not available in this sovereign Azure Local environment, which keeps the offer closer to an on-premises extension than a full sovereign public cloud. (techzine.eu) Microsoft’s message is that sovereignty can be built as an operating model on customer-owned infrastructure, with Azure as the control plane. The test now is whether governments and other regulated buyers treat that as a cloud alternative or as a bigger version of on-premises infrastructure. (blogs.microsoft.com) (techzine.eu)