ComputerWeekly

Counting cloud carbon cost an opportunity for new businesses


A common claim in technology is that moving to the cloud reduces carbon emissions. Hyperscale datacentres are typically more efficient than enterprise owned infrastructure, and cloud providers have invested heavily in renewable energy and operational efficiency.

The challenge however is to prove exactly how much carbon has been avoided. For organisations that report against sustainability targets, assess cloud migration programmes or offset residual emissions, there is an awkward reality. Namely, that carbon is surprisingly difficult to measure in the public cloud. 

A business cannot meaningfully offset a number it cannot confidently measure. And in many respects cloud sustainability has become an observability problem. 

Cloud teams have detailed metrics for performance, availability and cost. But carbon is hard because the information needed to calculate it sits across operational systems, energy networks and infrastructure platforms, depends on allocation methodologies and assumptions, and often cannot be independently verified by customers.  

This is an area where there is real opportunity for high potential businesses to start, scale and reach their full potential. That’s if they can develop a frontier digital technology that offers greater insight into the carbon cost of the public cloud. As organisations place greater emphasis on understanding the environmental impact of digital infrastructure, demand is likely to grow for solutions that can translate carbon data into operational insight. 

Offsetting a number you cannot see  

The first challenge that comes with offsetting a number that a business cannot see is attribution. In a traditional datacentre, for example, an organisation owns the infrastructure and can directly measure electricity consumption. 

In the public cloud, thousands of customers share the same infrastructure. Servers, storage systems, networking equipment and cooling systems are pooled across multiple workloads, tenants and applications. The relationship between a single workload and the energy required to support it is therefore far less obvious, and hinders how businesses can assess their productivity, performance and overall costs.  

Most cloud providers offer sustainability dashboards, but these typically provide aggregated monthly estimates rather than detailed operational data. An organisation may know its cloud estate generated a certain volume of emissions during a reporting period, but not necessarily which applications or business activities were responsible. This can make it difficult for business leaders to make informed decisions on where changes can be made, and the likely outcome of such decisions that pertain to efficiency, cost and output.  

To build a more accurate picture, organisations need to combine workload data, energy data, environmental attributes such as carbon intensity, and information about the underlying infrastructure. Only by linking these different sources can they understand how digital activity translates into environmental impact. Without this level of insight, carbon accounting becomes an exercise in allocation and estimation rather than direct measurement.  

The baseline problem 

What makes it difficult to develop a solution to this problem, however, is that carbon reduction requires comparison against a baseline. Yet establishing a stable baseline in the public cloud is far from straightforward and creates a significant challenge for organisations and opportunity for high potential businesses developing frontier technologies to address it. 

Cloud environments are constantly changing. Workloads scale automatically. Applications are modernised. Data volumes increase. AI services are adopted. Infrastructure shifts between regions. The environment being measured rarely stays still long enough to make straightforward comparisons. 

Location adds another layer of complexity. Carbon intensity can vary dramatically between cloud regions and even between different times of day within the same region. 

Consider a data processing workload scheduled overnight in a region supplied primarily by wind or hydroelectric generation. Run the same workload during peak demand in a region dependent on gas fired generation and the resulting emissions may be considerably higher. Nothing has changed within the application itself. What has changed is the electricity system that supports it. 

From measurement to optimisation 

The real opportunity is not simply better carbon reporting. It is bringing operational, energy and environmental data together in a way that allows organisations to make better decisions. 

When workload data is combined with energy consumption, carbon intensity information and infrastructure characteristics, organisations gain a clearer picture of environmental impact. That creates opportunities to shift workloads to lower carbon regions, improve utilisation and make more informed operational decisions. 

More importantly, it moves sustainability beyond annual reporting and into day-to-day operations. Emerging approaches such as carbon-aware computing and carbon-aware workload scheduling move flexible workloads to locations or times where lower carbon electricity is available. This is where the challenge becomes less about accounting and more about innovation. Improving environmental performance and bringing information together could enable new tools, services and business models that help operators understand the carbon impact of digital infrastructure and its wider demands on energy systems and local resources. 

The public cloud is often more efficient than traditional enterprise infrastructure, but efficiency is not the same as transparency. Until workloads can be linked more clearly to energy use, environmental conditions and underlying hardware, organisations will continue to struggle to prove whether they are genuinely reducing emissions or simply reporting them differently. 

Frontier technologies as a foundation for change  

Addressing these challenges will require new ways of combining operational, infrastructure and environmental data. Encouragingly, high potential businesses across the UK are already beginning to tackle different aspects of the problem. 

Emerging innovators such as Greenpixie, Climatiq and Open Climate Fix demonstrate how frontier technologies that span cloud, data and AI can be applied to improve carbon transparency, support more informed decision-making and accelerate decarbonisation across the economy. From measuring the impact of cloud and AI workloads, to embedding carbon intelligence into digital systems and optimising renewable energy networks, these organisations highlight the growing opportunity for innovation at the intersection of digital technology and sustainability. 

Digital Catapult’s role is to help high potential businesses bridge the gap between technological innovation and industry adoption. By working with industry partners to identify challenges, validate solutions and create routes to commercialisation, we help promising businesses demonstrate value in real-world environments and scale their impact. We do this by working across high-growth sectors including clean energy, defence, security and the creative industries.  

As organisations place greater emphasis on understanding the environmental impact of their digital infrastructure, the ability to measure, manage and optimise carbon performance is likely to become an increasingly important source of competitive advantage.

The public cloud has transformed how organisations consume computing resources. The next challenge is ensuring that environmental performance can be understood with the same level of visibility and confidence as cost, availability and security. 

Luke Sperrin is head of clean energy industries at Digital Catapult  



Source link