Munich dumps proprietary for open source (again) – but no Linux desktop

Munich is moving parts of its municipal IT stack from proprietary products to open-source equivalents, but it is not bringing back LiMux, the Debian/Kubuntu-based Linux desktop that once made the city the most cited example of open-source adoption in government. Dr. Laura Dornheim, Munich's chief digital officer, laid out the current strategy in a keynote at Open Source Summit Europe in Prague, saying the earlier Linux desktop effort ended for political rather than technical reasons.
The concrete swaps are infrastructure and collaboration tools. The city is replacing VMware with Proxmox for virtualization, moving from Atlassian's Jira and Confluence to OpenProject and XWiki, and adopting PostgreSQL in place of Oracle databases. Munich expects the two infrastructure and collaboration migrations alone to save roughly €1.5 million or more each year, and Dornheim put the overall open-source shift at around a million euros in annual savings in her remarks.
That ordering is deliberate. Rather than treating a desktop operating system as the centerpiece, Munich is replacing individual layers of proprietary infrastructure where it can make a practical financial case. The distinction matters for anyone who remembers the first campaign, because the new program is narrower in scope but arguably more durable in structure.
The history is worth recalling precisely, since it is often misremembered. Munich's council voted in May 2003 to build its own Linux distribution as Windows NT 4.0 approached end of support. Rollout began in earnest in 2006, and by December 2013 the city had completed the migration with more than 14,800 desktops running LiMux. A European Commission Open Source Observatory case study reported savings of about €11 million from dropping Windows.
LiMux became the leading proof point that a large city could run Linux desktops and open-source office software at scale. It also exposed how hard it is to change a sprawling government IT environment when Microsoft Office documents, legacy workflows, PDF handling and user expectations remain tied to proprietary systems.
The political weather changed after Munich's 2014 municipal election. By November 2017 the council had voted to return its computers to Windows with a target of 2020, and the office-suite transition from LibreOffice to Microsoft Office was not finished until 2023. Dornheim's explanation for the reversal was blunt: a new government had secured an agreement from Microsoft to place the headquarters of Microsoft Germany in Munich. Her broader lesson was that open source in the public sector rarely fails on technical grounds, but often fails when it is not aligned with the political agenda and lacks durable political backing.
That framing shapes the current approach. Munich's IT strategy now aims to make open-source software work well while keeping the political support needed to sustain it — which, in practice, means tying projects to budgets, public services and the ability of different parts of the administration to share tools. Digital sovereignty is a factor, but Dornheim indicated it is not the primary driver; cost savings and control are. She also noted that a cloud transition to Azure without sufficient assurance from Microsoft on GDPR compliance would oblige the city to scrutinize who could use each tool, a management complication that does not strengthen the case for staying with Microsoft.
Supporting structures have changed too. Munich established an Open Source Program Office in 2024 to coordinate the work, and its public software portal organizes activity into three categories: use existing open-source software, improve it, and publish software the city develops itself. The city has also embraced the "public money, public code" principle, meaning software built with public funding should where appropriate be available for others to inspect, use and improve — a commitment applied to newly acquired or developed software even while the earlier Windows migration was still running. Dornheim became the city's CDO in September 2022 and has not treated that policy as a mandate for another administration-wide operating-system overhaul.
Munich is now a participant in open-source development rather than only a customer choosing between licenses. One example is DAVe, its application for documenting and analyzing traffic data, in use since 2021 and adopted by Wolfsburg. Dornheim argued that such matches between cities should be routine rather than accidental, and that a structure spanning the whole public sector is needed to make them happen. The city also runs MUCGPT, a municipal AI application giving employees access to large language model capabilities for tasks such as writing and proofreading emails, and an Open Source Sabbatical program that supports developers working on open-source projects, with the aim of improving and funding components used in municipal services.
A likely trade-off in this layered approach is that savings and control accumulate gradually and unevenly, since each migration carries its own integration, training and support burden. Virtualization, project tracking, wikis and databases are areas where open-source alternatives are mature and well understood, which supports Munich's cost case. Desktop Linux remains the harder problem precisely because it touches document compatibility and everyday workflows across an entire workforce — the friction that dogged the first campaign — so leaving it aside removes the most visible symbol of open-source commitment while also removing its heaviest operational risk.
Why it matters: Public-sector technology teams weighing similar migrations get a concrete, budgeted template — swap proprietary infrastructure components where open-source equivalents are mature, and leave the desktop question for later. The figures Munich cites, roughly €1.5 million or more annually from the infrastructure and collaboration moves, give procurement and platform teams a benchmark to test against their own license renewals. The political reading is the durable lesson for product and IT leaders: in government settings, technical viability alone rarely decides whether an open-source program survives a change of administration, so aligning projects with budget, service and interoperability priorities is part of the engineering work, not a separate communications task.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Open source in the public sector doesn’t fail because of technical issues. It fails because it’s not aligned with the political agenda