Cloud repatriation: Why the future is hybrid
What’s driving cloud repatriation, and what does it teach us about the future of IT infrastructure?
Updated May 15, 2025)
IT infrastructure is always evolving. Trends swing like a pendulum, from cloud-skeptical to cloud-first, and everywhere in between. In recent years, cloud repatriation continues to gain traction. It’s clear that organizations are reevaluating how to best balance performance, security, control, and long-term cost.
Cloud repatriation refers to moving workloads from public cloud environments back on-premises. It doesn’t always mean leaving the cloud, but is more about finding the right balance, often a hybrid setup.
Household brand names like Droptox and 37signals have publicly documented their journeys to bring workloads back on prem.
Dropbox saved nearly $75 million by moving away from cloud solutions in 2018.
In 2023, 37signals’ CTO David Heinmeier Hansson made headlines for saying the cloud isn’t the only way. Recently, Hansson shared he anticipates saving an additional $1.3 million annually, which will lower overall infrastructure costs to under $1 million per year.
How? By migrating the last remaining 18 megabytes of data from AWS to Pure Storage arrays that cost less than $200,000 a year to operate. By summer 2025, Hanssen anticipates they’ll be able to delete their AWS account altogether.
“Cloud can be a good choice in certain circumstances,” Hansson notes. “But the industry pulled a fast one convincing everyone it’s the only way. No wonder you see cloud vendors and ads and PR everywhere. There’s so much money in convincing everyone that owning your own hardware is impossible…”.
Realistically, not every business can replicate cost savings on this scale. But these examples highlight a growing willingness to question the cloud status quo.
The Platformonomics Repatriation Index suggests repatriation is likely more of a rebalancing than a mass movement. Instead, we’re seeing more nuanced hybrid cloud strategies emerge.
Adopting Site Reliability Engineering (SRE) practices may correlate with public cloud to on-prem repatriation. SRE principles, focused on operational excellence and reliability, can influence organizations to reevaluate their infrastructure strategy and consider repatriating workloads.
It’s not just about cost savings. Multiple factors are at work here.
Here are the key drivers:
Cost optimization: Cloud costs are unpredictable. They’re also often higher than initially projected. 451 Research and IDC rank cost as the top reason for repatriating workloads. Add to that, Statista estimates 30% of total cloud spending is wasted on egress fees and overprovisioning.
Control and security: Data sovereignty, visibility, and breach response are harder to manage and contain in public cloud environments. The attack surface is often a known unknown, which is a gamble many organizations in regulated industries cannot take. IDC’s 2018 survey first highlighted enhancing security and compliance as a primary driver for repatriation. Breaches and outages from AWS, Azure, and GCP have only increased scrutiny.
Regulatory compliance: Some industries need certain workloads and data to stay on-prem to meet compliance requirements. Common examples include finance, government, defense, and critical infrastructure. Disaster recovery mandates also drive local infrastructure needs. Managing data on-premises ensures compliance with data protection standards.
Performance and latency: Some workloads demand low-latency or uninterrupted access that the public cloud doesn’t always guarantee. On-premises solutions can offer more predictable performance. This becomes especially critical for isolated networks or those without an internet connection.
Vendor lock-in: Relying on a single cloud provider can limit flexibility and introduce a single point of failure. Reducing dependency on a single provider for critical workloads, especially an external provider, is becoming a strategic priority.
A recent Microsoft Active Directory community poll confirms these drivers are rooted in real-world realities.
93% of respondents operate today in a hybrid environment with most workloads still on-prem.
36% expect to stay that way permanently,
31% estimate it’ll take 10+ years to move fully to the cloud.
The future isn’t cloud or on-prem, it’s hybrid. Businesses are learning to align workloads with the environment that best fits them. Some applications benefit from the scalability and flexibility of a cloud migration. Others are better served by on-prem for control, compliance, or technical reasons.
This rebalancing isn’t always loud or dramatic. Unlike viral blog posts from SaaS founders, most companies are making these decisions quietly, guided by CFOs and CTOs. Their focus is on total cost of ownership, security, resilience, and long-term sustainability.
The goal for organizations today isn’t to go full-cloud or be on-prem purists. It’s to build a smart infrastructure that’s secure, cost-effective, and helps get the job done. Reassessing workload placement in light of operational priorities, security concerns, and financial realities is just good governance.
If cloud repatriation highlights one thing, it’s the need to stay flexible. We’ve all been told the cloud is the future. And it may well be for many organizations, but hybrid IT isn’t a compromise. It’s a strategic choice that offers the best of both worlds.
Over the next 10 to 15 years, most organizations will turn to hybrid infrastructure to meet broader security, business, and financial goals.