Over the past five years, one of the most consistently tracked figures in the UK business technology sector has been the flight from public cloud.
Barclays' 2021 CIO survey revealed that 43% of enterprises plan to shift workloads away from public cloud. By 2024, that had grown to 83%.
Research for Pulsant in 2025 found that 87% of UK businesses planned to repatriate data away from the public cloud within the next two years.
These figures make it clear that the unstoppable force which has powered public cloud has hit an immovable object.
The force was the idea that limitless compute can be a cheap commodity afforded by hyperscale public cloud and delivered flexibly on a ‘pay as you go’ basis.
The object was the reality of cost – specifically cost at enterprise scale. The need to improve both the ROI of Infrastructure as a Service, and IaaS cloud cost control, has led to a keen interest in the budgetary impact of getting the right workload in the right place.
The Rising Challenge of IaaS Cost Control
Repatriation began with businesses realising that technology strategy needs to fit the workloads organisations already have in place – not the other way around.
The pay-as-you-go approach of public cloud works for varied workloads that demand flexibility, especially if the overall demand remains low.
But for workloads that scale significantly or behave predictably (and so do not need to pay a premium for occasional flexibility), that model can erode margins.
The problem has been made worse as businesses have had to penetrate complex pricing models that reduce cost visibility and make forecasting unnecessarily difficult.
This complexity is often by design. A fear of performance bottlenecks leads to overprovisioning, where companies pay for a "safety buffer" of compute power that never actually does anything.
This is then complicated by the actual operating costs. For example, "cold" storage tiers might offer lower monthly rates, but then the high fees associated with data retrieval and API calls (where one application requests data or services from another) can quickly offset any savings if the data is accessed more often than anticipated.
In effect, the very flexibility that was a selling point of public cloud penalises businesses that do not use it.
And then, at the point when these issues come to a head and a business needs to leave the public cloud, it is hit with costly egress and data transfer fees. Moving data into the cloud is typically free but shifting it between regions or out to the internet is expensive.
It is typically at this point that a business explores alternative options and encounters private, managed cloud or Infrastructure as a Service (IaaS).
How IaaS Improves Cloud Cost Control
Public cloud is defined as a third-party provider offering data storage, transit, and processing, using the public internet, to more than one organisation.
A private cloud delivered as Infrastructure as a Service (IaaS), sees the storage, processing and transit of data provided by a third party to a single business, typically on privately held, managed infrastructure.
It is the difference between arriving at the airport on holiday and facing the choice of a public bus or booking a private limo.
This is not solely a consideration of the quality of experience. Managed IaaS infrastructure enables businesses to regain financial control through predictable cloud costs and better workload placement.
Attempts to anticipate long-term demands on the architecture will create a much more reliable foundation – and keep the infrastructure aligned to the business.
Typically, private IaaS cost optimisation is built on fixed, predictable pricing models compared to variable public cloud billing. Providers can offer dedicated resources to a client and offer greater control over configuration.
Critically, this can all be offered in sovereign environments, hosted in UK data centres, with connecting network infrastructure that keeps data onshore.
Having identified the right kind of workload to make migration worthwhile, how does a business know if it has such a workload?
IaaS in practice - selecting the ‘right’ workload
When it comes to optimising workloads, there are four main candidates for managed IaaS:
- Steady-state workloads with predictable demand – there is little point in paying the cloud tax of hyperscale elasticity for a potential spike in demand that is unlikely to arise.
- Latency sensitivity – some applications demand exceptionally low latency, and businesses risk a ‘lag tax’ if they need to travel distances to centralised hyperscale facilities. These are better suited to an edge architecture, processed closer to the point of collection.
- Workloads with large or frequent data transfers – this transit will not only rack up expensive fees but also expose the company to large egress fees should it find a better alternative than legacy public cloud.
- Sensitive workloads – using data that is legally required to stay on shore.
Having now identified the workloads to shift away from public cloud, and indeed those that might need to stay there, the next question is how best to migrate to a balanced hybrid.
Moving workloads from public cloud to IaaS
The first consideration is to ‘rightsize’ workloads - ensuring that CPU, memory, and storage allocations match actual requirements rather than theoretical peaks.
Based on a foundation of actual data on workloads required, the business can balance cost and performance by putting the right workload in the right place.
Steady-state, predictable workloads can be colocated or on-premise. Low latency workloads can be kept close and, when needed, public cloud elasticity can be exploited for experimental projects.
This leads into improved sustainability. The infrastructure remains lean, cost-efficient, and capable of supporting long-term business growth without budget bloat.
With a real-time dashboard, a business can identify orphaned resources or the pricing anomalies that suggest something is not right.
Hybrid infrastructure options - what enterprises are learning
The endgame is a cloud cost optimisation strategy where businesses pick what is needed from a portfolio of public, private and IaaS environments. That secures the best performance whilst keeping financial control.
Our research found that of those planning to repatriate, more than half are considering a move to a private cloud provider (54%), whilst 38% are exploring an increased reliance on their own data centres, and 36% are assessing colocation.
For organisations shifting from "cloud-at-any-cost", the consistent theme is to keep control and ensure technology matches the business.
If you are looking to reduce unpredictable cloud spending or build a cost-efficient enterprise architecture that keeps you in control, please email us or make an enquiry to optimise your infrastructure with Pulsant IaaS.
