Executive Summary

The costs of migration away from Microsoft cannot be answered with a single figure, but rather arise through comparison of three cost curves: status quo Microsoft with documented steep growth dynamics, one-time migration costs for Open Source plus parallel operation, and ongoing operating costs of the Open Source stack. Reliable figures are available from several major European projects: Schleswig-Holstein shows the strongest economic reference with EUR 9 million in one-time investments for 30,000 workplaces and more than EUR 15 million in annual license savings. The French Gendarmerie migrated over 13 years with approximately EUR 2 million in annual direct savings. Munich (LiMux) demonstrates governance risks: the re-migration cost EUR 50–86 million. France (DINUM) is planning national scaling in 2026 with estimated EUR 1.5–3 billion over 5–7 years. Switzerland rolled out M365 broadly in 2025 and is currently examining Open Source alternatives in a Proof of Concept.

People

Topics

  • Digital Sovereignty
  • Open Source Migration
  • IT Costs and TCO
  • Administrative Digitalization
  • Cloud Strategy

Clarus Lead

Microsoft spending by public administrations is multiplying: The Swiss Federal Government spent CHF 42 million in 2009 but CHF 140 million in 2026 – a tripling in 15 years. Germany increased its software license budgets from EUR 771 million to over EUR 1.2 billion (+57%) between 2022–2023, with Microsoft products nearly quintupled. This cost dynamics creates political and economic pressure to act, forcing European authorities to develop Plan B strategies. Schleswig-Holstein and France demonstrate that structured migrations are economically amortizable – but only with clear governance, adequate funding, and multi-year political continuity. The greatest avoidable cost trap is federal fragmentation without central coordination: parallel custom solutions in cantons, federal states, and municipalities multiply efforts and sabotage economies of scale.

Detailed Summary

Cost Curve Comparison and Data Quality

The question of migration costs requires distinction between three competing financing scenarios. Status quo Microsoft is not a static model: Microsoft 365 functions as a subscription with structurally rising costs. The announcement of price increases for July 2026 (single-digit to low double-digit percentages) confirms this characteristic. For every 100,000 users, annual license cost jumps occur in the hundreds of millions range. Schleswig-Holstein calculates EUR 500 per person per year as current license savings when eliminating Microsoft 365 across Europe.

The available authority figures are heterogeneously defined: Some sources cite pure license costs, others Total Cost of Ownership (TCO) including personnel, hardware, and support. Munich's LiMux controversy illustrates this: the City of Munich calculated its Linux costs in 2012 at EUR 23 million, while a parallel HP study commissioned by Microsoft calculated EUR 60.7 million. The central difference lay not in accounting, but in assumptions about porting business applications. If all business procedures must be completely migrated to Linux, costs escalate considerably. If they can remain operational via terminal server, web, or transitional architecture, the calculation becomes considerably cheaper.

Schleswig-Holstein as Primary Reference Case

Schleswig-Holstein is currently the strongest publicly documented economic reference. The state government has reported the following concrete results since 2024: By the end of 2025, nearly 80% of workplaces (excluding the exempted tax administration) were working with LibreOffice. Almost 44,000 email mailboxes had been migrated to Open-Xchange. The government calculates the one-time migration investment for 2026 at EUR 9 million and the annual license savings from 2026 at more than EUR 15 million compared to Microsoft 365.

Mathematically, this yields a license-based payback in less than one year (EUR 9 million investment ÷ EUR 15 million annual savings). The per-capita license savings amounts to approximately EUR 500 per person per year based on 30,000 workplaces. Additionally, Schleswig-Holstein is funding 17 Open Source projects with nearly EUR 3 million.

Critical Assessment: These figures refer to the investment and savings framework for 2026 alone. They do not fully include: personnel and preparation costs from previous years (establishing coordination, pilots, organizational structures), costs of areas still bound to Microsoft (tax administration, Windows-only business applications), indirect costs of shared platform development via Dataport or North German administrative cooperation, operating costs for identity management and business procedures over the longer term. The significance lies not in completeness, but in the fact that for the first time a current European government project publicly names the concrete relationship between investment and savings.

International Reference Cases: Gendarmerie, Munich, France, Germany

The French Gendarmerie (GendBuntu) has gradually migrated over 100,000 workplaces since 2004/2008 to its own Ubuntu variant. Direct annual license savings are approximately EUR 2 million (EUR 20 per workplace per year). Reported TCO reduction is around 40%. The migration strategy was decisive: the Gendarmerie migrated applications before the operating system – first OpenOffice/Firefox on Windows, then Linux. This reduced user acceptance problems. However, this migration occurred in an era of classical Office and Windows licenses. Today's Microsoft 365 cloud stack is more expensive, which is why modern license savings are significantly higher.

Munich (LiMux, 2003–2017, re-migration from 2017) is a governance case study, not a cautionary tale. The City of Munich calculated its Linux costs in 2012 at EUR 23 million for 15,000 workplaces. A concurrent HP study (Microsoft commission) calculated EUR 60.7 million. The difference is primarily explained by differing business application porting assumptions. The actual damage did not arise from the migration itself, but from political reversal: the re-migration to Windows 2017–2022 cost EUR 50–86 million, of which EUR 29.9 million was pure licenses, EUR 24 million external consulting, EUR 14 million personnel costs, and EUR 13.4 million IT service providers. These costs would have been avoidable through clear political continuity across legislative periods.

France (DINUM 2026) announced on April 8, 2026, that its own government workplaces would switch from Windows to Linux and that all ministries submit reduction plans for non-European IT dependency. Eight domains are affected: operating system, collaboration, antivirus, AI, databases, virtualization, cloud, and networks. The frequently cited costs of EUR 1.5–3 billion over 5–7 years are analyst estimates, not confirmed budget figures. With up to 2.5 million workplaces, these would amount to EUR 600–1,200 per workplace. France frames the migration not as an isolated Linux project, but as an industry, security, and sovereignty policy program. Success depends on implementation discipline, ministerial coordination, and business procedures.

Germany (ZenDiS/openDesk) demonstrates an underfunding problem: ZenDiS was founded in 2022 with 9 employees (4 in development) and 2024 funding of EUR 19 million on an assignment basis (predominantly OpenCoDE). By February 2025, EUR 950,000 had been allocated for openDesk and EUR 550,000 for OpenCoDE. The Open Source Business Alliance demands EUR 30 million per year as a minimum. A 2024 budget remainder of EUR 34 million was released by the Budget Committee but, according to reports, was not disbursed. This illustrates that national ambitions without adequate base funding and binding rollout obligations come to nothing.

Cost Components and Success/Failure Patterns

Directly avoidable costs include: Microsoft 365 subscriptions, Windows client licenses, server/backend licenses (Exchange, SharePoint, Teams, Windows Server), Microsoft-specific training, and partially endpoint security licenses.

One-time migration costs typically distribute across: external consulting 25–40%, internal personnel costs 15–25%, hardware adjustments 5–10%, training 10–20%, business application adaptation/replacement 20–60% (largest cost block), migration tools/conversion 5–10%, parallel operation 5–10%.

Success factors are: migrate applications before operating system (Gendarmerie approach), secure political mandate across legislative periods, institutionalize Open Source Program Office (OSPO), pursue upstream strategy rather than fork proliferation, prioritize web-based workflows, make training a core budget item.

Failure patterns: fragmented federal structures without coordination, underfunding of OSS coordination (ZenDiS problem), test pilots without binding roadmap, underestimating business application complexity, lack of political continuity.

Swiss Case Calculation

A hypothetical scaling to the Swiss Federal Government (54,000 workplaces, analogous to Schleswig-Holstein) yields:

  • One-time migration investment: approximately CHF 16 million (analogous to Schleswig-Holstein EUR 9 million × scaling), with surcharges for multilingualism, federalism, and special tasks (VBS/EDA) CHF 24–32 million.
  • Annual license/cloud savings: approximately CHF 27 million (54,000 × CHF 500).
  • Calculated license-based payback: 1–2 years under favorable assumptions.
  • Net effect over 10 years (greatly simplified): CHF 240 million (10 × CHF 27 million − CHF 32 million).

Most important limitation: This calculation provides only an orientation. A complete TCO would need to separately show: business applications and Windows-only dependencies, identity and access management, SharePoint/OneDrive/Teams data migration, parallel operation Microsoft/Open Source, training and change management, procurement/security certification/audit, operating costs of the Open Source target stack, transition costs for VBS, EDA, and protected environments.

With reported Swiss Federal Microsoft spending of CHF 140 million per year, not the entire sum would be quickly substitutable. More realistic is a replaceable component (licenses and cloud) of CHF 20–40 million per year, depending on scope, hybrid operation, and business procedures. Savings would come from gradual reduction of rising subscription, cloud, and lock-in costs over several years, not from immediate Microsoft elimination.

City of Zurich (15,000 workplaces analogous to Schleswig-Holstein) would have: migration investment approximately CHF 5–8 million one-time, annual savings approximately CHF 7–8 million, license-based payback around 1 year under favorable assumptions. The city's current Microsoft framework agreement through 2030 comprises up to CHF 108 million (CHF 83 million administration, CHF 25 million