
Table of Contents
Cloud Repatriation: Evaluating the Shift from Public Cloud to On-Premise Solutions
- Updated on


There’s a lot of buzz going around recently about cloud repatriation, and whether or not the sky is falling. I’ve spent some time researching this, as many articles and videos I’ve come across focus mainly on financials to determine the future of public cloud. While I initially took that approach too, I believe there’s more to the story that’s being overlooked. As Nagios has always focused on on-premise monitoring, I think there’s a lot of nuances being missed in the discussion. Let’s dig into it.
What is Cloud Repatriation?
Companies that move away from public cloud infrastructure like Azure, AWS, GCP, and are instead moving to on-premise deployments, data center co-locations, or similar options are considered to be repatriating their infrastructure. For example, a company may decide that a few servers that are used to host an application for internal users doesn’t need to be in the cloud anymore. Those servers can be repatriated on-prem, or to a co-location datacenter.
The current assertion going around seems to be that companies are repatriating their systems in droves.
Key Drivers Behind the Cloud Repatriation Trend
Back in May of 2021, Sarah Wang and Martin Casado of Andreessen Horowitz wrote an article talking about the long-term costs of moving to the cloud, citing some industry-heavy hitters like Dropbox, CrowdStrike, and Zscaler. The article is best summed up by their description of the paradox they’re talking about.
“You’re crazy if you don’t start in the cloud; you’re crazy if you stay on it.” -Sarah Wang and Martin Casado, The Cost of Cloud, a Trillion Dollar Paradox
Financial Considerations: Cloud Costs vs. On-Premise Investments
Several people have commented on the latest financials for the big three cloud providers, noting that they’re still showing growth in the billions of US dollars. This suggests they’re doing well financially. But this only shows how they’re doing overall and doesn’t necessarily mean that their cloud services are providing that growth.
One of the biggest things overlooked in the financials discussion is the huge AI bubble. Companies large and small are building large language model-based applications for customers. The AI bubble may be driving a lot of that growth. For example, Statista reported that AI grew to 184 billion US dollars in 2024.
If there is downward pressure on public clouds’ financials, it’s likely being led by companies that have their own reasons for migrating away from the public cloud space. One of those reasons might be the cost of trying to fit a square peg in a round hole.
Something I mentioned in a previous webinar was that many years ago, in the hype to move to Azure and AWS and GCP, people lost sight of their use cases. Infrastructure as a Service (IaaS) is where a lot of companies ended up running a lot of systems that would’ve been better off being run in a virtual private server (VPS) environment.
The trick with companies coming to their own conclusion that they’re not on the right platform is that the speed at which those companies leave is a relative trickle, compared to the companies that jump onboard when there’s a lot of hype involved. In this particular case, I think it’s safe to say that the AI hype train is large enough to overpower any downward pressure on a public cloud’s financials due to companies migrating away.
Assessing Whether Cloud Repatriation Aligns with Your IT Strategy
Ah, the tag line everyone loves to hear. “It depends.” I can’t tell you what it depends on, but I can tell you what it doesn’t depend on. What other people are doing. You, the person reading this article, you have an environment, you have servers that you’re responsible for. We’re professionals. You don’t need herd mentality to tell you where to stand up your next server.
Ultimately the various *aaS products that public cloud providers offer are a tool. VPS is also a tool. On-premise is also a tool. You know what the requirements are for your existing environment, you know what the requirements are for the next few systems you’ve been asked to stand up. Look at the tools available and pick the one that makes the most sense.
Final Thoughts: Weighing the Pros and Cons of Cloud Repatriation
Well, is the sky really falling or not? No. The big 3 aren’t likely to fail any time soon. Based on some of the other articles I’ve read, there are enough companies out there moving back to on-prem or colocations that there isn’t much overlap in the example companies provided by those articles. Best I can tell, there are a large number of professionals out there who have reevaluated what tools they have, and are picking a different tool to get the job done.
My belief is that there is at least a fairly strong move away from public clouds, but AI came along at just the right time to give a big boost, and more than cover any losses from companies slowly migrating their systems away. This lines up with the typical zerg rush* we see when a new tool is hyped up, and its slow decline as people find that the tool is really for a different audience.
* Zerg rush: A term originating from the video game Starcraft, where a stronger and/or better equipped force is defeated by a much larger force of comically weaker and/or poorly equipped units.