Lenovo expands Hybrid AI Factory for sovereign AI deployments

Lenovo is expanding its Hybrid AI Factory portfolio with Microsoft, combining Lenovo infrastructure and services with Azure Local and Microsoft Foundry Local. The offering is designed to help organizations run AI closer to their data and operations while retaining greater control over workloads, security and governance.
- Lenovo is expanding its Hybrid AI Factory with Azure Local and Microsoft Foundry Local.
- The portfolio supports local inference, conversational RAG, agentic AI and vision AI.
- New GPU options target AI inferencing, video processing, larger models and higher inference demand.
- Confidential computing and selected data protection capabilities support sovereign AI deployments.
- Lenovo is adding Sovereign AI Discover advisory services and expanded Azure Local services.
AI designed to run across local and cloud environments
Lenovo says organizations adopting AI must manage regulatory requirements, inference costs and application responsiveness while deciding where workloads should run. Those locations can include a desktop, an enterprise datacenter or the cloud.
The expanded Lenovo Hybrid AI Factory brings together Lenovo ThinkAgile, Azure Local, Foundry Local and Lenovo Services. According to Lenovo, the portfolio supports local inference, conversational RAG, agentic AI and vision AI in environments operated and trusted by customers.
Azure Local gives teams access to familiar Azure tools and experiences for managing workloads across hybrid cloud environments. Organizations can run workloads locally or in Azure depending on their business needs, according to Lenovo.
Lenovo ThinkAgile provides the infrastructure foundation for virtualization, private and hybrid cloud, and AI. Its ThinkAgile MX Series combines AI-ready hyperconverged infrastructure and disaggregated architectures with Azure Local.
More infrastructure options for different AI workloads
Lenovo ThinkAgile MX650a V4 systems featuring NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs are positioned for organizations that want to bring AI workloads closer to their data, applications and users.
Lenovo is also adding support for NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs on the ThinkAgile MX650a V4. The company says this option is intended for AI inferencing and video processing, helping customers balance performance, power and cost.
For larger models or higher inference demand, the Lenovo ThinkSystem SR680a V4 with an NVIDIA HGX B300 platform and eight GPUs provides another option, with access to models from the Foundry Local catalog.
Lenovo says these configurations allow customers to select capacity for current workloads while planning for larger models or higher inference demand. Lenovo Top Choice Express is intended to provide streamlined access to selected configurable infrastructure with faster fulfillment than traditional configure-to-order processes. Lenovo AI Express combines validated, right-sized AI infrastructure with software and services options to help organizations move from AI pilots to production.
Security, sovereignty and continuity controls
Lenovo frames sovereign AI as control over where data and workloads run, who can access them, and how they are operated and governed. Customers can choose hyperconverged or disaggregated architectures and operate connected or disconnected environments where supported.
Disaggregated deployments allow compute and storage to scale independently. With disaggregated Azure Local architectures, supported ThinkSystem DM, DG and DS Series external SAN storage configurations can help organizations expand storage separately from compute.
Depending on the selected platform and configuration, data protection capabilities include ransomware protection, post-quantum cryptography data protection, immutable retention and recovery. Lenovo is also introducing support for confidential computing with Azure Local on ThinkAgile MX Series solutions. The company says hardware-enforced isolation helps protect sensitive data while it is being processed.
Confidential VMs and confidential GPUs are intended to support deployments where privacy, resiliency, regulation and operational control are important. Lenovo says these capabilities can help organizations bring AI models into environments they operate and trust.
Services and workloads beyond AI
Lenovo’s expanded Hybrid AI Factory Services cover advisory, implementation and managed services for Microsoft Azure Local, including new Azure Foundry Local capabilities. Lenovo has obtained the Microsoft Azure Local for Hybrid and Sovereign Private Cloud Specialization, which recognizes capabilities across Azure Local assessment, architecture, deployment, workload migration and operational support.
The company is also introducing Sovereign AI Discover, an advisory engagement for assessing readiness, defining a posture for priority use cases and building an evidence-based roadmap for sovereign AI.
The same infrastructure can support virtual desktops, Windows Server workloads, data services and developer environments on customer-owned infrastructure. Azure Virtual Desktop Hybrid on Azure Local supports Windows 11 Enterprise multi-session, autoscale and Start VM on Connect. Lenovo also introduced the ThinkSystem SR450 V4, powered by Intel Xeon 6 processors, as a Windows Server option for branch and distributed IT environments.
Lenovo cited Gartner research published in October 2026, which found that 77% of heads of infrastructure and operations consider data sovereignty essential to their objectives, while 55% globally believe current technology sovereignty practices adequately preserve ownership and control. The expanded portfolio addresses that concern by giving organizations options to keep AI, data and applications in local environments, extend workloads to Azure when needed, and scale infrastructure as requirements change.
Lenovo: Bring sovereign, secure AI to where your business runs


