Rethinking Cloud Repatriation Strategies in 2025

Retarus’ Uwe Geuss offers commentary on rethinking cloud repatriation strategies in 2025. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

Global IT spending is set to reach US$5.61 trillion in 2025, a nearly 10 percent increase over the previous year, and enterprise leaders are facing a new challenge: do more with less. As budgets grow, so does the scrutiny on where that money is going. CIOs and CISOs are being asked to trim excess spending from IT budgets while still delivering on innovation and resilience.

One of the most significant factors in the debate over IT budgets is cloud infrastructure, which is facing growing pushback amid concerns about increased spending. 84 percent of enterprises cite managing cloud spend as a top challenge, and many organizations have reconsidered moving back to on-premises systems in pursuit of perceived cost savings. Others, like Fortune 500 companies Geico and Dropbox, have already returned to legacy systems earlier this year.

But what happens when decision-makers question whether cloud as a whole is too costly, and demand a return to on-premises infrastructure? IT leaders must be ready to explain the critical risks and hidden costs of total cloud repatriation before making the case for the emerging third way in infrastructure: bimodal IT.

The Risks of Total Cloud Repatriation

The unknown and underestimated costs of returning to on-prem systems wholesale are significant. Repatriating entire cloud systems to solve one pain point can introduce multiple others.

In addition to the upfront capital investment for new hardware, specialized employees and infrastructure, organizations must account for the substantial operational expenditure involved. Cloud providers have been able to absorb the costs of operation and maintenance and have also gained years of experience in this field. Abandoning the cloud could mean abandoning these capabilities to manage expenses.

Leaving the cloud also means leaving its high-availability service model, which keeps operations running smoothly and reliably. It’s impossible to put a price tag on the disruption prevented by this model. Suffice it to say that without cloud providers’ vast networks of redundant data centers and sophisticated disaster recovery solutions, businesses returning to on-prem infrastructure may only realize how much they’ve been saving in the cloud when they incur the costs of an outage or slowdown.

The same warning applies to compliance and security. Cloud providers alleviate the resource and cost-intensive processes of securing and maintaining the certifications businesses need to work in highly regulated sectors. They also lead in building robust security frameworks that companies can rely on to prevent breaches that could devastate their bottom line. These responsibilities – and their respective costs – fall back onto the business when moving back to on-prem infrastructure, which could quickly negate any short-term cost savings made by leaving the cloud.

The migration process alone could prove costly and disruptive. Data transfers, compliance overheads, and potential downtime can quickly cause customer dissatisfaction and revenue loss. In 2024, 75% percent of cloud repatriation projects failed to achieve their intended cost savings within 18 months, mainly due to unforeseen complexities.

However, the most significant downside of cloud repatriation extends beyond the bottom line and could have far greater repercussions for a business’s success. For over two decades, cloud infrastructure has been celebrated as a catalyst for innovation and agility, and is seen as essential to creating a future-proof business. From deploying AI models to integrating third-party APIs, many modern tools are built for the cloud and designed to leverage its scalability and flexibility.

Cloud repatriation could slow these digital transformation initiatives. After over a decade of operating in cloud-native environments, reverting to on-prem systems could stall momentum and risk a shortage of skilled administrators and employees with on-prem expertise.

These downsides, along with the need to hire or retrain staff for legacy systems, could erode any short-term financial advantages. In the long run, wholesale cloud repatriation will likely leave organizations struggling to maintain, secure, and update aging infrastructure on their dime.

Finding the Right Mix of On-Premises & Cloud 

The downsides of on-premises infrastructure do not mean that cloud is completely faultless, or that legacy systems are inherently bad. The cloud repatriation trend is part of a broader evolution in infrastructure strategy that has taken place over the past two decades in enterprise IT, where organizations experiment and learn which approach works best for their specific needs.

Indeed, while there is a trend of migration of selected workloads back to on-premises environments, cloud infrastructure remains advantageous for many use cases.

For example, cloud migration is still prevalent in messaging as organizations seek its advantages. Cloud infrastructure offers the availability and reliability to manage high messaging volumes during peak times, while reputable cloud providers can provide various industry-specific certifications to ensure organizations remain compliant with the necessary regulations. The cloud is a factor in better cost management and is a more suitable choice for this particular use case.

Retarus was one of the first organizations to recognize the value of the cloud for enterprise messaging. That’s why our messaging services are based on cloud infrastructure but are designed to integrate into cloud, on-premises and hybrid infrastructures. This ensures a reliable, secure and cost-effective experience for organizations that depend on fax, email and SMS for business-critical communications.

The best IT teams keep their proverbial eggs in multiple baskets. While it is fair to seek out ways to reduce costs from the cloud, a bimodal approach ensures that teams can deploy workloads in the environments where they make the most sense.