Why are companies leaving the cloud?

Over the last year or so, the future of hyperscale cloud has been hot on the IT agenda. Having dominated the infrastructure world for the past decade, and pulling in record profits, the undisputed reign of hyperscale cloud is coming into question.

Many organizations are seeking to exit, motivated by unmet expectations including spiraling costs, limited support, increasingly unmanageable complexity, security concerns and an overall lack of control. Some issues are more pertinent than others, but what’s clear is that companies are seeking out alternative solutions that mean a partial, if not total, shift away from hyperscale cloud.

As businesses weigh up the benefits of on-prem, colo and infrastructure as a service deployments and especially hybrid architectures, what are the reasons for their hyperscale cloud exit?

Unexpected costs 

The escalating costs associated with hyperscale cloud certainly made headlines in 2024. It’s understandable then that a lack of transparency when it comes to infrastructure costs is one of the biggest issues faced by businesses today. In fact, 42% of businesses using hyperscale cloud struggle to predict their monthly cloud bill, 28% have received an unexpectedly large bill for cloud services, and 82% end up wasting more than 10% of their overall cloud spend on hyperscale infrastructure.

For many, it all starts from a seemingly innocent choice made at the very beginning of a company’s cloud journey. To get customers signed up, hyperscale cloud providers offer difficult-to-refuse free credits that allow a company to get up and running quickly. Soon enough a business is designing its entire IT infrastructure around a hyperscaler’s specific, proprietary products. Problems arise when the free credits run out and the bills start. Customers are left unable to migrate to more cost-effective solutions without completely rearchitecting or re-engineering. And many IT managers don’t have experience carrying out migration projects of this type.  

On top of that comes the need for extra security and support from the cloud giant, which sit on a seemingly endless sliding scale of usage-based costs. Cloud resource usage optimization can help, but ultimately businesses are still confronted with growing fees. As recent examples show, the most significant savings are to be had by either integrating alternative sources of compute or migrating away from hyperscale cloud completely.

Security concerns

Public hyperscale cloud environments are prime targets for security challenges, so it is unsurprising to know that 61% of businesses experienced a cloud security incident in 2024. Worryingly, 21% of which resulted in a data breach.

It was only in January 2024 that we saw a major data leak in Microsoft Azure which left hundreds of executive accounts compromised, and up to 97,000 Microsoft Exchange servers vulnerable to attack. It’s one of many examples that have influenced companies to manage their own infrastructure, data and security protocols, particularly for organizations dealing with sensitive data.

The inherent security risks brought by public cloud are a product of its shared responsibility model. In these environments, there is always a risk that another business residing on an organization’s server will make a security-compromising mistake. And because most data breaches result from human error not hardware failure, the risk is always present.

33% of organizations cite security issues as their primary motivation for leaving the public cloud. And, given that sharing hardware leaves businesses vulnerable to the mistakes of others, mitigating these security risks by leaving public cloud environments seems like a natural step. 

A lack of control, and dwindling support

Many assume that the prestige of a big-name cloud label brings excellent levels of support, but for many organizations, this couldn’t be further from the truth. In reality, there are negligible levels of support for those only paying the minimum subscription fee, with personalized one on one help available only to those paying premiums on top of an already hefty bill. A problem may be big for a business, but if it isn’t critical for your cloud supplier, it’s unlikely to be addressed with any urgency.

Break-fix-only support is standard, and in most cases, comes via a link to a generic knowledge base article. Little support for when things go wrong, paired with the fact a business has limited control over its public cloud environment, means serious ramifications can be on the cards. Whilst going on-prem doesn’t fix the support issue, it does mean a business will have complete control over managing and troubleshooting its technology, providing it’s got a specialized team. 

That said, for businesses wanting the support of a hosting provider, there are good hyperscale cloud alternatives out there that offer tailored and reliable support. Including the option for bare metal server hosting, with specialised providers able to personally work with businesses to overcome challenges in real time.

Application-specific requirements

Many organizations assume that public hyperscale cloud environments are a fix-all to every type of workload. The reality is, this perception has been born out of clever marketing campaigns, and not all applications are best suited to the cloud. For example, hyperscale cloud products have been designed for modern applications and many enterprise workloads that currently sit in the cloud should probably never have been put there in the first place.

A few years ago, when cloud adoption was the hottest trend in IT, it wasn’t unusual to see enterprises move their legacy applications to the cloud en masse. What these companies failed to realize was that these legacy applications weren’t set up to benefit from the features offered by hyperscale cloud environments, like containerization and Kubernetes. What was left in the shadows were businesses subject to inefficiencies, increased costs and reduced performance over time - the result of incompatibilities between their applications and their hyperscale cloud environment.

The lesson here for all businesses is that unless there is a real need to refactor applications for the cloud, chances are they’re going to be better off being run on dedicated servers instead.

Going all in - is it the only choice?

Many businesses are realizing that a nuanced, flexible approach to infrastructure is the winning formula, with many avoiding the pitfalls of overinvesting in a single technology. As a result, we’re seeing hybrid infrastructure deployments becoming increasingly commonplace. Complex, diverse workloads need hybrid IT models to support, blending the best of compute and storage in hyperscale cloud whilst also utilizing on-premises and colocation deployments.

Public cloud isn’t a panacea, but instead a technology that can be an incredibly effective tool when used in the right circumstances. Making infrastructure choices on a workload-specific basis enables organizations to optimize performance, cost and increase redundancy by avoiding the single point of failure by going all-in on public hyperscale cloud. 

Cloud certainly serves a purpose, but the way businesses have historically been pushed towards cloud services has been inefficient and ineffective. With a renewed, nuanced approach to infrastructure, businesses are moving their technology back on-prem, into colocation or hosted with an IaaS provider with great success. But for most, it means adopting a hybrid approach that combines the best compute types for each workload.