Cloud repatriation: why businesses are reshoring their workloads

Faced with billion-euro budgets getting out of hand and a level of overreliance that has become nearly impossible to undo, more and more businesses are considering moving their workloads away from the public cloud.

Cloud repatriation is mainly driven by predictable, long-term applications, which are the ones affected by skyrocketing budgets due to ongoing usage-based and data traffic fees, as well as managed services premiums. And then there are requirements to be complied with regarding data sovereignty, transparency and the ability to carry out regulatory checks. In highly regulated industries or in the public sector especially, they are more easily and reliably met through on-premise or colocation solutions. Technical factors such as latency, data gravitation and performance predictability are additional drivers for cloud repatriation.

The total cost of ownership (TCO) also plays a key role in this decision. In public cloud environments, the TCO is often underestimated, as processing power and memory aren’t the only cost factors: data traffic, observability, platform services, idle resources, possible price hikes and, especially, exit costs are often incorrectly priced. Over the space of several years and at constant load profiles, this can lead to major deviations from the original forecast. What often happens is that players operating long-term applications with a stable load profile end up paying more for using the cloud than they thought.

Digital sovereignty and the trap of ‘sovereignty washing’

If there is one topic closely related to reshoring, it’s digital sovereignty. Digital sovereignty refers to the ability to operate mission-critical IT systems independently from a legal, technical and business standpoint and to switch providers if necessary and to carry through this decision. ‘Sovereignty washing’ describes a situation whereby pledges to that effect remain empty words, while dependency remains because of a proprietary interface or overseas jurisdiction. Cloud repatriation is one of several options to ensure effective portability and clear responsibilities.

Things become particularly challenging when companies seeking sovereignty-compliant solutions specifically turn to those market-dominant providers that they already rely on heavily. This does not seem advisable as this only conceals the dependency instead of removing it. True sovereignty exists if, and only if, swapping providers remains a feasible option and players have full control over their data at any point in time.

Strategies, workloads and hybrid infrastructure in practice

But what does repatriation look like in practice? Well, this strongly depends on the size of the organisation. Medium-sized businesses tend to have a pragmatic, targeted approach and only reshore a small number of clearly defined applications that are subject to high cost pressure or require a high level of sovereignty. Major companies, on the other hand, tend to proceed in waves, working with graduated governance, domain roadmaps and accompanying sourcing contracts to comply with regulatory requirements and remove any complex webs of dependency in an orderly manner. Despite the many paths to get there, the aim remains the same: having greater control over one’s infrastructure.

In this respect, hybrid architectures are well suited, as they combine both sovereignty and flexibility. Critical or latency-sensitive components operate on-premise or in colocation centres, while variable peak loads and global services can be flexibly sourced from the cloud. Portability by design is key and ensured through stable deployment procedures, infrastructure as code, constant monitoring and coordinated security rules. Switching from cloud to cloud is supported by the provisions of the new Data Act. In hybrid architectures, costs can also be kept down thanks to used software in BYOL models.

The systems particularly well suited to reshoring are long-term, predictable ones with a high level of inbound/outbound data such as ERP, DMS or PLM solutions as well as data-intensive platforms with large volumes of internal data traffic. Licensed applications with operating costs that are easier to manage on-premise are also ideal candidates, in which case users can also resort to the second-hand market, which is perfectly legal in the EU. What usually remains in the cloud are highly variable or globally spread workloads, content distribution and short-term AI/ML prototyping, as long as this does not jeopardise data protection and cost control.

Avoiding typical mistakes and regaining freedom

Nevertheless, there are a number of traps to be avoided when moving away from the cloud. Users often underestimate the time and budget needed to repatriate data, especially in the case of large databases with no upstream seeding or parallelisation. Another common mistake is relying on a mere ‘lift-and-shift’ without planning any replatforming steps, meaning that any inefficiencies are replicated on-premise. Structural differences as well as their impact on licensing must also be considered during migration. Additionally, in many cases, there are no clear exit rules regarding data formats, deletion confirmations and support hand-over nor any reliable tests regarding identities, interfaces and network segments in the target operating model.

Another issue is that skills have often been lost after years of overreliance on the cloud. This can only be offset by having clarity when reversing the trend. What works best is gradually building up know-how, with teams establishing classical PlatformOps based on storage, network and virtualisation, as well as modern practices like GitOps, infrastructure as code and policy as code in parallel. At the same time, it is worth consolidating expertise in procurement and licences to responsibly use all business options, including used software – a perfectly legal option. Against this backdrop, IT providers can become real enablers if they proactively help with portability, exit design and transparency and provide well-structured knowledge.

Conclusion

For all these reasons, cloud repatriation is set to be part and parcel of any professional multisourcing strategy in the long run. Price cycles may vary, but the structural drivers, i.e. sovereignty, cost control, reversibility and transparent compliance, are here to stay. In 2026, modern on-premise IT systems are API driven, highly automated and based on KPIs. They are based on HCI and fabric infrastructure, self-service portals, declarative deployments, integrated FinOps as well as zero-trust principles from the data centre to the edge. LizenzDirekt has always been both a procurement specialist and an independent adviser ensuring that IT systems meet the needs of organisations while complying with legislation and European values. As a dealer in used software and a fervent supporter of this market in the EU, we help businesses with interoperable implementation across software companies and make sure licenses are audit-proof. Our aim at the end of the day: to guarantee technical sovereignty, profitability and compliance in any repatriation project.