Where should a workload actually run? 

For UK enterprises, the answer more often depends on the demands of individual applications and the infrastructure available to support them. Some workloads benefit from cloud scalability, while others may be better suited to infrastructure hosted on business premises. 

The UK Government's Business Data Survey 2026 has illustrated the range of approaches already in use. Among businesses handling digitised data, 31% reported using public cloud services, 27% used servers on business premises, and 19% used private cloud infrastructure. These categories can overlap, reflecting the different environments businesses use to store and process information. 

Speaking during Scaling AI Responsibly: Energy, Efficiency & Innovation in Data Centres at Tech Show London, Matthew Baynes, Vice President Secure Power UK&I at Schneider Electric, highlighted the challenge of designing infrastructure around changing technology requirements. 

“Right sizing, the engineering brief, has never been more important,” said Baynes. 

His observation applies to the choice between hybrid cloud and on-premises infrastructure. Overestimating requirements can leave organisations paying for unused capacity, while underestimating them can restrict future growth. The workload's requirements, and how they might change, should inform where it runs. 

Hybrid Cloud Offers Flexibility With Added Complexity 

Hybrid cloud connects cloud services with infrastructure hosted on-premises, allowing organisations to use both environments. 

For businesses with changing demand, cloud resources can provide additional capacity without requiring equivalent investment in physical infrastructure. Hybrid approaches can also support gradual modernisation, allowing existing applications to remain on-premises while newer services move into the cloud. 

However, connecting environments introduces dependencies. 

The National Cyber Security Centre (NCSC) advises organisations adopting hybrid cloud to understand where data is stored, how it moves between environments, and whether connectivity can support the additional demand. 

Availability also requires attention. If an application depends on both cloud and on-premises components, a connectivity failure could affect the entire service unless resilience has been designed into its architecture. 

Hybrid cloud provides more workload-placement options, but those options introduce additional management responsibilities. 

On-Premises Infrastructure Puts Control in the Hands of the Business 

On-premises infrastructure remains relevant where organisations need direct control over equipment and operating conditions. 

Predictable workloads, applications with demanding latency requirements, and systems that are difficult to migrate may be suitable candidates. Local infrastructure can also offer greater control over how sensitive information is stored and processed, although data security depends on the safeguards implemented. 

That control carries responsibility. 

Businesses operating their own infrastructure must plan for hardware replacement, capacity expansion, and maintenance. Where they operate their own facilities, they must also account for power, cooling and physical resilience. 

The operational question that remains is whether the organisation has the expertise and resources to maintain those systems throughout their lifecycle. 

For some workloads, retaining existing infrastructure may be practical. For others, the investment required to maintain or expand it could make cloud services more attractive. 

Compare Costs Across the Workload's Lifetime 

Cloud and on-premises costs are often discussed in terms of operating expenditure versus capital investment. That comparison can overlook costs that emerge after deployment. 

Cloud expenditure depends on consumption, storage, and networking requirements. Costs can increase as workloads scale or data moves between services. 

On-premises infrastructure requires upfront hardware investment and ongoing expenditure on maintenance, facilities, and staffing. Equipment replacement and unused capacity also affect the longer-term calculation. 

Provider dependency deserves consideration. 

The Competition and Markets Authority's cloud market investigation reported that UK organisations spent approximately £9 billion on cloud services in 2023. Its provisional findings identified technical and commercial barriers that could make switching providers or using multiple cloud platforms more difficult. 

For enterprises, the implication is that cloud flexibility does not automatically guarantee portability. 

A meaningful comparison should consider the expected lifetime of the workload, how its requirements might change and the cost of moving it later. 

Operational Complexity Can Determine Whether Hybrid Works 

Hybrid cloud may provide an answer to workload placement while making day-to-day operations more demanding. 

IT teams need visibility across environments that may use different monitoring systems and security controls. Identity management becomes particularly important when users and applications require access across cloud and on-premises infrastructure. 

The NCSC advises organisations to adapt their access controls so that compromising one environment does not automatically give an attacker privileged access to another. 

Operational integration also matters during incidents. Teams need to understand which services depend on each other, where failures originate and who is responsible for responding. 

A related concern emerged during Automating Operations Beyond the Build at Tech Show London, where Chris Markstein, Data Centre Lead at Inductive Automation, discussed the importance of infrastructure systems exchanging information. 

“If it can't talk to the other systems in your facility, it doesn't belong on your floor,” said Markstein. 

Although Markstein was discussing data centre automation, the principle also applies to hybrid infrastructure. When systems operate across different environments, fragmented monitoring and disconnected operational information can make it harder to identify problems and coordinate a response. 

Which Approach Fits the Workload? 

The following considerations provide a starting point for assessing infrastructure placement. 

Consider Hybrid cloud when...	Consider on-premises when...
Demand varies significantly	Workloads are relatively predictable
Additional capacity is needed quickly	Direct control over infrastructure is a priority
Different workloads have different hosting requirements	Existing applications are closely tied to local systems
Modernisation needs to happen gradually	Latency or operating requirements favour local hosting
Teams can manage connected environments	Teams can support the physical infrastructure throughout its lifecycle

 

These are considerations rather than fixed rules. A predictable workload may still be more economical in the cloud, while sensitive data can be hosted securely in either environment when appropriate controls are in place. 

The infrastructure decision begins with the workload, but it also depends on what the organisation can support over time. 

For UK enterprises, choosing between hybrid and on-premises infrastructure means weighing the value of flexibility against the cost of managing it. The most suitable environment is the one that meets the workload's requirements while remaining financially and operationally sustainable.