From NAS OS to Governed Private AI Platform
Synology’s next-generation DiskStation Manager (DSM) is defined as an evolution from a network-attached storage interface into a governed private AI platform that turns on-premise AI data, logs, and documents into a secure knowledge base for human users and autonomous agents. Built on more than 20 years of operating system development, 14 million systems shipped, and over 400 exabytes of data under management, DSM’s evolution aims to keep data ownership inside the enterprise while still enabling modern AI capabilities. Synology positions this as a response to growing volumes of unstructured content, tighter digital sovereignty demands, and security expectations that make public cloud alone insufficient. As Philip Wong, Synology’s Chairman and CEO, explained, “Enterprise AI adoption is no longer the challenge, data control is,” framing the platform as a way to adopt AI without relinquishing control of sensitive information.
Private AI Workspaces and Local Inference Infrastructure
Synology is turning DSM into a private AI platform that can process text, audio, images, and scanned documents on-premise. Synology Office Suite now includes an AI Assistant for content generation, email summarisation and replies, natural-language formula queries, live translation, and meeting notes, giving teams a private generative AI workspace. Beyond productivity, DSM will offer speech-to-text, OCR, image captioning, and embedding to convert existing content into an internal knowledge graph without defaulting to public cloud services. Model deployment is designed to be flexible: DSM will support OpenAI API-compatible large language models, enabling both self-hosted and cloud-based options, and Synology plans built-in local LLMs for simpler setups. The company is also preparing GPU-enabled RackStation 26-series systems and AI Station appliances for local inference and multi-GPU clusters, so enterprises can keep on-premise AI data and intellectual property inside their own racks instead of external data centers.
DSM Agent and the Rise of Agentic IT Workflows
At the core of DSM evolution is DSM Agent, a built-in AI agent that turns AI from an add-on tool into an engine for agentic workflows. Initially, DSM Agent will act as an on-demand assistant for troubleshooting, health checks, service monitoring, and proactive alerts across the Synology ecosystem. Over time it is meant to perform more autonomous operations, such as backup validation, gap detection, and guided restoration. For security teams, DSM Agent will help analyse suspicious logins, anomalous file activities, and log indicators. Governance is woven through the design using guardrails, data loss prevention, user and group permissions, and fine-grained controls over skills, tools, DSM packages, and files. DSM will also expose MCP and CLI access so third-party enterprise AI tools and other agents, including models like Claude, can connect under the same permission model, integrating AI into daily IT operations without weakening security controls.
Fleet Management NAS and Clustered Enterprise Storage AI
To support enterprise-wide on-premise AI data strategies, Synology is extending DSM from single-appliance control to fleet management NAS capabilities. Cluster Manager will let administrators treat multiple Synology systems as one pool, improving resource utilisation, performance consistency, and deployment flexibility across sites. Under the hood, DSM introduces a container-based storage infrastructure that isolates workloads such as file services, Synology Drive, and MailPlus, which in turn supports smoother workload migration, storage quality-of-service, and load balancing. Active Insight’s Mass Deployment will allow batch rollouts and configuration of many units, with centralised monitoring of health, updates, and AI-related services. According to Synology executives, the new DSM is “built for both AI and enterprise demands, enabling private AI workflows with full governance, fleet-scale management, and the security controls IT teams need for regulation and compliance requirements,” positioning DSM as an enterprise storage AI control plane rather than a single-box NAS UI.
Security, Compliance, and the Case for On-Premise AI Data
Security and compliance sit at the centre of Synology’s DSM roadmap, which is expected to add FIPS 140-3 level cryptographic compliance alongside expanded governance controls. By keeping inference, knowledge bases, and system telemetry on-site, organisations can process on-premise AI data without exposing confidential records or system logs to shared cloud environments. DSM’s layered permissions—from user groups and packages to AI tools and actions—are designed to make audits easier and reduce the risk of data leakage through autonomous agents. The platform’s approach aligns with enterprises that prefer a single capital expenditure on infrastructure rather than recurring public cloud fees and bandwidth charges, while still gaining AI-assisted automation. As internal documents, metrics, and logs turn into inputs for autonomous workflows, DSM’s security posture and governance rails will likely be the deciding factor for IT leaders weighing private AI platform adoption against expanding cloud dependence.
