Data sovereignty means a company keeps control over where business data live, who can access them, and how they are exported or deleted. With cloud ERP, operations shift to the vendor. Responsibility for compliance and control stays with the company. Equating “cloud in Germany” with data sovereignty falls short.

What data sovereignty covers in practice

Four dimensions belong together:

1. Location. In which data center do production and backup data live? Which legal jurisdictions and subcontractors are involved?
2. Access. Who at the vendor has technical access? How are roles, logging, and emergency access governed?
3. Export. Can data be exported completely and in a machine-readable format, including history and documents?
4. Termination. What happens after the contract ends to data, backups, and proof of deletion?

Without these four points, “data sovereignty” stays a marketing phrase. With them, it becomes auditable.

Why location alone is not enough

GDPR-compliant hosting in Germany or the EU is a baseline for many mid-market projects, not a differentiator. What decides is contract, operating model, and exit. A server in Frankfurt with opaque subprocessors, missing export, and unclear deletion deadlines is not a sovereign solution. Conversely, a cleanly documented cloud model with tested portability can give more control than an on-premise system without backup drills and without documentation.

Cloud vs. on-premise is one dimension of the debate, not the whole. Digital sovereignty comes from deliberate decisions about dependencies, not from the label of the operating model.

Which questions must be asked before the contract

Put these check questions into the RFP and contract review:

• Where do primary data, replicas, and backups sit geographically and organizationally?
• Which subcontractors have access, and how are changes communicated?
• In which formats and within which deadlines does a full export happen?
• What does data handover on termination cost and how long does it take?
• How long may data still exist after the contract ends, and who confirms deletion?
• Can you restore backups independently of the vendor dashboard?

Test export and restore with a sample before you go live. Paper clauses without a sample are risky. More under ERP exit strategy.

How on-premise and hybrid fit in

On-premise gives maximum infrastructure control and requires an internal or external operations team for patches, monitoring, and recovery. Hybrid models make sense when ERP and data stay portable: for example sensitive core systems internally, less critical services in the cloud. What matters is that the data model and interfaces do not block switching between operating models.

Whoever chooses cloud should secure portability and switchability contractually and technically. Whoever chooses on-premise should assess operations competence and backup discipline realistically. Both can be sovereign. Both can also create lock-in when export and documentation are missing.

What this means for ERP selection

Data sovereignty belongs in every ERP RFP, on equal footing with functions and price. Leadership should clarify which dependencies are acceptable and which are not. IT should be able to evidence location, access, export, and termination. Open platforms make control easier because source code and data stay with the company. Nuclos supports cloud and on-premise with the same feature set. The right variant follows from compliance, resources, and risk profile, not from the buzzword of the season.

Data sovereignty is control, not autarky. Whoever masters contract, operations, and export stays able to act, even when the vendor changes.