Cloud computing has become ubiquitous in modern IT systems. First gaining prominence in the early 2000s with the rise of AWS, organizations have adopted cloud computing due to the allure of improved scalability, flexibility, and cost efficiency. However, cloud repatriation is a growing movement that is beginning to disrupt this narrative. As organizations migrate workloads back from public clouds to on-premises or hybrid environments, the mainframe is reemerging as a critical player in this transformation. This shift is not merely a technical adjustment. It signals a strategic reevaluation of where enterprise computing truly belongs, and how the mainframe’s role is evolving in a cloud-saturated world.

The Case for Cloud Computing

The rise of cloud computing was driven by its perceived usefulness to achieve higher scalability, reduce costs, and improve the speed of deploying systems and applications.

Cloud computing offers dynamic, elastic scalability, allowing organizations to seamlessly adjust resources up or down as demand fluctuates, without the burden of maintaining costly, and sometimes underutilized hardware. This model also transforms traditional IT spending by replacing large upfront capital expenditures (CapEx) with a more flexible operational expenditure (OpEx) approach. This is alluring to organizations, enabling them to pay only for the computing power they actually consume when they consume it.

And the marketing siren song of deploying quicker is another selling point of cloud computing. Cloud providers tout the ability to rapidly provision infrastructure, enabling faster time-to-market for new applications.

In short, moving to the cloud offered a cost-effective means to offload IT management, allowing companies to focus on their core competencies. At least that was the working theory.

The Challenges of Cloud Adoption

Despite the cloud's advantages, many organizations are now reconsidering their cloud strategies. Several factors are contributing to this trend of cloud repatriation.

While the cloud's pay-as-you-go model sounds attractive, many organizations are finding that costs can spiral out of control. Without proper governance, usage spikes can result in unexpected bills. Additionally, sustained workloads — those running 24/7 — are often more cost-effective on-premises, as the cloud's dynamic pricing may not provide enough cost benefits.

Additionally, some workloads, particularly those that are data-intensive or latency-sensitive, may not perform as well in a cloud environment. Network latencies, bandwidth constraints, and the geographical location of cloud data centers can introduce performance bottlenecks, making on-premises or hybrid solutions more attractive. For example, financial services data processing and real-time automation control systems are probably not well-suited for cloud environments due to their high dependency on access to physical processors and low-latency network transactions.

Data sovereignty and security are also drivers for platform consideration. With increasing regulations around data privacy (e.g., European Union’s General Data Protection Regulation (GDPR), California Consumer Privacy Act (CCPA)), companies need to ensure that sensitive data is stored and processed in compliance with local laws. Cloud repatriation allows businesses to regain control over their data and ensure it is stored within specific geographical boundaries, providing peace of mind from a compliance standpoint.

Many organizations initially embraced the cloud for its flexibility, but over time, they have found themselves tied to specific vendors. This vendor lock-in makes it difficult to switch providers or move workloads across different platforms, leading to a desire to repatriate data and applications to more neutral environments.

Back to the Mainframe

Repatriating workloads from the public cloud back to the mainframe can yield significant cost savings, especially for organizations managing high-volume, transaction-intensive workloads. While cloud platforms offer attractive entry points, the cumulative costs of storage, data egress, and continuous compute consumption can escalate rapidly over time.

Mainframes, by contrast, provide predictable pricing models and exceptional resource efficiency, allowing businesses to optimize utilization without the variable expenses associated with cloud services. For enterprises that already maintain mainframe infrastructure, repatriation also maximizes existing investments, eliminating redundant cloud costs and consolidating operations under a more financially sustainable model.

From a performance standpoint, the mainframe remains unmatched in its ability to handle massive workloads with low latency and exceptional reliability. Modern mainframes are engineered for high throughput, capable of processing billions of transactions daily without degradation in speed or responsiveness. By moving critical workloads back on-premises, organizations can reduce network latency caused by data transfer between cloud regions or across hybrid architectures. This localized computing approach ensures faster access to data and more consistent performance, particularly valuable for industries like finance, healthcare, and retail, where real-time processing is essential.

Security is another area where mainframes continue to demonstrate superiority. With built-in encryption, secure partitioning, and centralized access control, mainframes offer a level of hardware-based protection that’s difficult to replicate in a public cloud environment. Repatriating sensitive workloads (such as customer data, financial transactions, or intellectual property) to the mainframe allows enterprises to maintain tighter control over their security perimeter and compliance posture. Moreover, keeping critical data on-premises minimizes exposure to third-party vulnerabilities, while enabling organizations to adhere more easily to strict data residency and regulatory requirements.

As organizations bring critical workloads back to the mainframe, the integration of quantum-safe cryptography is becoming an essential part of long-term security strategies. Mainframes are already at the forefront of encryption technology, and with the rise of quantum computing, the platform is evolving to protect data against future quantum-based threats. By adopting quantum-safe algorithms designed to resist the computational power of quantum attacks, enterprises can ensure that sensitive information remains secure for decades to come. This proactive approach strengthens the mainframe’s role as the most trusted foundation for mission-critical workloads in a rapidly changing threat landscape.

The ideal approach is to run workloads on the mainframe that require high performance, reliability, and security at scale because these are the things at which that mainframe excels. Mainframes are purpose-built for processing massive volumes of transactions, maintaining data integrity, and supporting mission-critical operations where downtime or latency simply isn’t acceptable.

The following chart outlines various types of workloads that are ideal for mainframes:

Repatriation and the Hybrid Multi-Cloud

Repatriation does not necessarily mean a complete abandonment of the cloud. In fact, many organizations are opting for hybrid or multi-cloud architectures that provide the best of both worlds. By maintaining a combination of on-premises infrastructure and cloud services, businesses can optimize their workloads based on cost, performance, and security requirements.

Hybrid cloud, as its name suggests, is a hybrid approach allowing organizations to run critical, latency-sensitive workloads on-premises while leveraging the cloud for less demanding, scalable tasks. This approach also can provide better control over data governance and compliance.

Instead of relying on a single cloud provider, a multi-cloud approach allows organizations to distribute their workloads across multiple cloud platforms. This strategy avoids vendor lock-in and enhances redundancy, improving resilience and uptime.

Both are tactics used by organizations to hedge their bets on a wholesale move to a cloud infrastructure. And both can include the mainframe as an essential, core component.

In a modern hybrid cloud approach, the mainframe can serve as the core processing and data integrity engine, anchoring the enterprise IT ecosystem. While the cloud provides agility, scalability, and rapid innovation for front-end applications and customer experiences, the mainframe ensures that critical workloads (e.g., transaction processing, data management, compliance operations, etc.) run with unmatched reliability and security.

By integrating mainframes with cloud environments through APIs, containers, and orchestration tools, organizations can modernize their operations without compromising performance or governance. This balanced approach allows enterprises to leverage the best of both worlds: the resilience and efficiency of mainframes alongside the flexibility and innovation of the cloud.

Strategic Considerations for Repatriation

Cloud repatriation is not a decision to be taken lightly. Companies need to carefully weigh the pros and cons. The Total Cost of Ownership (TCO) of any solution is an important criterion.

Moving workloads on-premises requires an upfront investment in hardware, software, and personnel. Businesses need to evaluate whether these capital expenditures are justified compared to ongoing cloud costs. When it comes to the mainframe, most organizations that adopted workload in the cloud did not completely abandon the mainframe. This means that the hardware outlay is not as substantial as an initial investment but may require investment in additional processors, memory, and storage as workload is repatriated to the mainframe.

Another consideration is assuring appropriate personnel and resources are available. Repatriating workloads means managing them internally again. This requires specialized talent such as system administrators, network engineers, and database administrators. Of course, even organizations that are fully implemented in the cloud have not completely abandoned staffing these skilled professionals. Repatriating workloads, however, will cause a shift in the expected tasks at which these professionals need to excel.

Finally, not all applications are suitable for on-premises environments. Modern, cloud-native applications designed with microservices and containerization in mind may not easily migrate back to traditional infrastructure without significant re-architecture.

Is Cloud Repatriation Here to Stay?

The cloud isn’t disappearing, but cloud repatriation serves as a clear signal that no single model can meet every organization’s needs. As businesses grow and their IT environments mature, priorities such as cost control, performance optimization, and data governance continue to evolve. Repatriation provides a strategic way to realign workloads, ensuring that each runs in the most efficient and secure environment possible.

While it may not be the ideal path for every enterprise, the repatriation trend highlights a broader truth: flexibility and adaptability are the cornerstones of modern IT strategy. For many organizations, the future lies in a balanced ecosystem that blends on-premises systems (including mainframes), cloud platforms, and hybrid architectures with each selected and configured to meet specific operational goals.

In essence, cloud repatriation is less about abandoning the cloud and more about achieving equilibrium. Whether an organization remains cloud-first, moves workloads back to the mainframe, or adopts a hybrid approach, long-term success depends on aligning technology choices with business priorities, performance demands, and strategic outcomes. By thoughtfully positioning each workload where it delivers the most value, enterprises can build a more resilient, efficient, and future-ready IT ecosystem.

To learn more about mainframe modernization and hybrid cloud, visit SHARE'd Knowledge.

Craig S. Mullins is president and principal consultant of Mullins Consulting, Inc. and  an in-demand analyst, author, and speaker. He has over four decades of IT experience in all facets of database systems development and mainframe management. He has written several books on database administration, Db2, and IBM Z. Visit his website, www.mullinsconsulting.com.