Cloud infrastructure was once evaluated primarily on three things: performance, scalability and cost.

That equation is changing.

As businesses move more workloads into the cloud, the infrastructure supporting those applications is consuming more energy, requiring more physical resources and contributing to a larger environmental footprint. At the same time, AI workloads are increasing demand for high-performance computing, making the relationship between digital growth and physical infrastructure impossible to ignore.

This is turning sustainable cloud infrastructure from a largely environmental discussion into a business one.

The issue is not simply whether an organisation can reduce its carbon footprint. It is whether the organisation can build and operate cloud environments that use energy, compute, storage and infrastructure efficiently while continuing to meet performance and availability requirements.

That distinction matters because inefficient cloud infrastructure creates more than environmental costs.

It can create financial waste, capacity constraints, operational complexity and long-term infrastructure risk.

Cloud Growth Still Depends on Physical Infrastructure

One reason cloud sustainability has been underestimated is the abstraction created by cloud platforms.

Businesses see virtual machines, managed databases, serverless functions and APIs. The physical infrastructure disappears behind the service interface.

But cloud computing still depends on servers, networking equipment, storage systems, cooling infrastructure, electricity and data centres.

The International Energy Agency estimates that data centres consumed around 415 TWh of electricity globally in 2024, representing about 1.5% of global electricity consumption. The IEA expects data-centre electricity demand to more than double by 2030, with AI playing a significant role in that growth.

For businesses, this creates a strategic consideration.

Every additional workload placed in the cloud ultimately depends on physical infrastructure somewhere.

The sustainability question therefore becomes closely connected to how efficiently those workloads are designed and operated.

AI Is Making Cloud Sustainability More Difficult

Traditional enterprise applications and AI workloads have very different infrastructure profiles.

A conventional business application may use relatively predictable compute resources. AI workloads can require significantly more intensive processing, particularly during model training and certain inference workloads.

The IEA estimates that electricity consumption from data centres could reach approximately 945 TWh by 2030, with AI identified as a major driver of the increase.

This does not mean businesses should avoid AI.

It means AI infrastructure decisions need to become more deliberate.

Running a model continuously when it is only needed periodically, storing unnecessarily large datasets, using oversized compute instances or moving data repeatedly between environments can increase both cloud costs and resource consumption.

The sustainability and FinOps conversations are therefore beginning to overlap.

The infrastructure that wastes energy often wastes money too.

Sustainable Cloud Is Not Simply About Choosing a Green Provider

A common misconception is that cloud sustainability can be solved by selecting a provider with renewable-energy commitments.

Provider-level sustainability matters.

But it is only part of the equation.

A highly efficient data centre can still host an inefficient workload.

Consider an application that runs oversized compute instances, stores duplicate datasets, performs unnecessary processing and retains inactive resources indefinitely.

Moving that application to a more energy-efficient facility may improve its footprint, but it does not eliminate the underlying waste.

Sustainable cloud infrastructure therefore needs to operate at multiple levels:

Data centre efficiency → cloud platform efficiency → workload efficiency → application efficiency

Businesses have meaningful influence over the last two.

That is where technology architecture becomes important.

Cloud Waste Has a Sustainability Cost

Cloud waste is usually discussed as a financial problem.

Unused compute. Idle databases. Oversized instances. Orphaned storage. Unnecessary data transfer.

But every unnecessary resource also represents physical infrastructure being used to support something that is not creating value.

This makes cloud optimisation relevant to sustainability.

Rightsizing workloads, removing unused resources, improving application efficiency and automating resource shutdowns can reduce cloud expenditure while also reducing resource consumption.

That creates an unusual alignment between two objectives that businesses sometimes treat separately.

The cloud environment that costs less to operate efficiently can also be the cloud environment that consumes fewer resources.

This does not mean cost and sustainability will always point to the same architecture.

They can conflict.

But there is considerable overlap.

Application Architecture Determines Infrastructure Efficiency

Cloud sustainability cannot be separated from software architecture.

An application that performs unnecessary database queries, repeatedly transfers large datasets or keeps resources running continuously will consume more infrastructure than an application designed around efficient workloads.

This is why cloud optimisation should not be left entirely to infrastructure teams.

Application developers influence resource consumption through architectural decisions.

Caching strategies, data-access patterns, API design, event-driven processing, storage policies and workload scheduling can all affect how much infrastructure an application requires.

A poorly designed application can therefore remain inefficient even when the underlying cloud environment is well managed.

Sustainability begins to become an engineering concern.

Serverless Does Not Automatically Mean Sustainable

Serverless computing is sometimes presented as a more efficient cloud model because businesses pay for execution rather than maintaining continuously running servers.

That can be true for certain workloads.

But serverless does not automatically make an application sustainable.

Poorly optimised functions can execute unnecessarily. Excessive invocation can increase resource consumption. Inefficient architecture can generate large numbers of downstream operations.

The same principle applies to containers, managed services and virtual machines.

The technology abstraction does not determine efficiency by itself.

Workload behaviour does.

Businesses therefore need to evaluate cloud architecture based on actual resource utilisation rather than assuming that a particular cloud technology is inherently greener.

Data Has Its Own Sustainability Footprint

Storage can feel almost free compared with compute.

That perception encourages organisations to keep everything.

Old backups remain indefinitely. Duplicate datasets accumulate. Temporary files become permanent. Logs are stored far longer than operationally necessary.

At enterprise scale, these decisions matter.

Storage infrastructure requires physical media, power, cooling and replication.

Data movement also consumes resources.

An architecture that constantly moves large datasets between regions or cloud providers may introduce both financial and energy costs that are invisible when teams focus only on compute pricing.

Data lifecycle management therefore becomes part of sustainable cloud architecture.

Businesses should know what data needs to be retained, where it needs to reside, how often it needs to be accessed and when it can be archived or deleted.

Carbon-Aware Computing Is Becoming More Relevant

Another emerging concept is carbon-aware computing.

Instead of treating infrastructure location and timing as fixed decisions, workloads can sometimes be scheduled according to factors such as electricity availability, grid carbon intensity or renewable-energy conditions.

Not every enterprise workload can operate this way.

Real-time customer transactions cannot simply wait for a cleaner energy window. Critical systems have strict availability requirements.

But certain workloads are more flexible.

Batch analytics, large-scale data processing, model training, testing and other non-time-sensitive workloads may offer opportunities for smarter scheduling.

This introduces another layer to workload management.

The question becomes not only:

“Where is compute cheapest?”

but potentially:

“When and where can this workload run most efficiently while meeting its requirements?”

Sustainable Cloud Infrastructure Needs Better Visibility

Businesses cannot manage what they cannot measure.

This is particularly difficult for sustainability because carbon impact is often less visible than cloud expenditure.

A finance team can see a monthly infrastructure bill.

An engineering team can see CPU utilisation.

But understanding the environmental impact of a particular application may require combining cloud usage data, infrastructure information, workload behaviour and energy-related data.

This is why sustainability needs stronger integration with existing cloud management and FinOps practices.

Organisations can begin by establishing visibility into:

• Compute utilisation and idle capacity
• Storage growth and retention
• Network and data-transfer patterns
• Resource provisioning and utilisation
• Workload-level cloud costs
• Energy and carbon information where reliable data is available

The objective is not to create another dashboard for executives.

It is to identify where infrastructure is being consumed without sufficient business value.

Sustainability Is Becoming an Architecture Constraint

Cloud architecture traditionally balances availability, performance, scalability, security and cost.

Sustainability is increasingly entering that decision framework.

For example, an organisation may have several ways to deploy an analytics workload.

One architecture might provide maximum performance but require substantially more compute. Another might process data more efficiently but take longer. A third might reduce resource consumption while introducing architectural complexity.

The correct decision depends on the business requirement.

The point is that sustainability is becoming part of the trade-off.

This is similar to security.

Businesses once treated security as something added after the application was designed. Mature organisations now incorporate security into architecture from the beginning.

Sustainability is moving in a similar direction.

It is becoming something architects need to consider before infrastructure is provisioned rather than after the cloud bill arrives.

Regulation and Reporting Are Adding Pressure

The sustainability conversation is also moving closer to corporate reporting and governance.

Large organisations increasingly face expectations around environmental reporting, supply-chain transparency and sustainability commitments. Technology infrastructure can therefore become part of a broader corporate sustainability picture.

This is particularly relevant for enterprises working with large customers, regulated industries or public-sector organisations where environmental requirements may influence procurement decisions.

The technology team may not own the company’s sustainability strategy.

But it increasingly contributes to it.

Cloud architecture, data-centre choices, infrastructure efficiency and software design can all influence the organisation’s broader environmental profile.

That makes sustainable cloud infrastructure a cross-functional issue involving technology, finance, operations, procurement and leadership.

The Business Case Is Stronger When Sustainability and Efficiency Align

Sustainability initiatives sometimes struggle when they are presented only as environmental commitments.

Cloud infrastructure provides a more practical business argument.

Reducing idle resources can lower costs.

Improving application efficiency can reduce infrastructure requirements.

Better data lifecycle management can control storage growth.

Smarter workload scheduling can improve resource utilisation.

Modernising inefficient applications can reduce both technical debt and infrastructure consumption.

The business case becomes much stronger when sustainability is connected to operational efficiency.

The goal is not to sacrifice performance simply to reduce energy use.

It is to stop paying for infrastructure that does not create corresponding value.

What Businesses Should Do Differently

Organisations do not need to redesign their entire cloud environment overnight.

A more practical approach is to identify where resource consumption and business value are misaligned.

This can begin with several actions:

• Measure before optimising: Establish visibility into workload utilisation, storage, data transfer and cloud consumption.
• Right-size infrastructure: Match compute and storage resources to actual workload requirements rather than peak assumptions.
• Build sustainability into architecture reviews: Include resource efficiency alongside security, performance, resilience and cost.
• Modernise inefficient applications: Address architectural patterns that create unnecessary compute, storage or data movement.
• Apply lifecycle policies: Archive or remove data that no longer needs to remain in expensive, frequently accessed storage.
• Separate workload types: Treat real-time, batch, experimental and AI workloads differently rather than applying one infrastructure model to everything.

The objective is not to make every application as small as possible.

It is to make infrastructure proportional to the value the workload provides.

Sustainable Cloud Will Become Part of Technology Strategy

The next phase of cloud maturity will not be defined simply by how much infrastructure a business has migrated.

It will increasingly be defined by how intelligently that infrastructure is used.

AI will increase compute demand. Data volumes will continue growing. Cloud environments will become more distributed. Businesses will need greater control over infrastructure costs and resource consumption.

That makes efficiency strategically important.

Sustainable cloud infrastructure is therefore unlikely to remain a specialist sustainability initiative handled by a small environmental team.

It will increasingly become part of architecture, engineering, FinOps and infrastructure planning.

The strongest organisations will not treat sustainability as a trade-off against performance or growth.

They will look for ways to make the technology estate more efficient precisely because efficiency supports both.

How Verbat Technologies Helps Businesses

Building a more sustainable cloud environment often starts with understanding the applications running on it. Legacy architecture, inefficient integrations, unnecessary data movement and poorly sized workloads can all increase infrastructure consumption.

Verbat Technologies works across cloud solutions, application modernization, custom software development, data engineering, enterprise application integration, API development, DevOps, AI/ML and digital transformation to help businesses improve the technology foundations behind their digital operations.

The objective is not simply to move workloads into a “greener” cloud environment.

It is to build applications and infrastructure that use resources intelligently, scale according to actual demand and continue delivering business value without unnecessary consumption.

Cloud sustainability is ultimately becoming less about the colour of the data centre and more about the efficiency of everything running inside it.

The most sustainable workload may be the one a business never needs to run unnecessarily in the first place.