According to information published yesterday by Dagens industri, the Swedish National Bank (Riksbanken) has approved an ambitious AI roadmap that will see the institution build its own sovereign AI platform capable of processing highly sensitive and classified information. Rather than relying exclusively on external AI providers, the Riksbank will run AI models on its own infrastructure, enabling employees to securely use AI across critical financial operations while maintaining full control over data, governance, and security.
The initiative goes well beyond deploying AI assistants. The roadmap outlines a long-term vision in which AI agents will interact directly with the Riksbank’s internal systems and databases to support economic analysis, monitor financial exposures between institutions, analyse transaction flows and balance sheets, identify unusual patterns, and explain movements in key economic indicators by automatically retrieving relevant historical analyses. The platform will also assist employees with report preparation, document summarisation, quality assurance, and internal data analysis.
The Riksbank will continue allowing employees to use public AI services such as ChatGPT, Claude, Gemini, Perplexity and Microsoft Copilot when working with non-confidential information. However, all sensitive and classified workloads will be handled by the internally hosted AI environment, reflecting a growing enterprise trend toward separating public AI from secure, mission-critical AI infrastructure.
To support the initiative, the Swedish National Bank plans to invest approximately 15 million SEK in AI servers between 2026 and 2028, alongside annual operating costs and consulting investments. Despite the extensive AI programme, the Riksbank states that the objective is to improve operational efficiency rather than reduce headcount, noting that its workforce has continued to grow over the past decade.
The announcement represents another clear signal that enterprise AI is entering a new phase. For highly regulated organisations, the conversation is rapidly shifting away from giving employees access to AI chatbots and toward building sovereign AI platforms capable of running trusted AI agents securely within the organisation. As governments, financial institutions and critical infrastructure providers accelerate these investments, sovereign AI is increasingly becoming a strategic capability rather than simply another technology initiative.
Add a comment
