BP, Shell, and most other major energy companies have pledged to reach net zero, alongside a real reduction in the emissions associated with every unit of energy they sell. Even accounting for the pace of the energy transition, the world will keep relying on hydrocarbon-based products for decades yet, which makes every genuine lever for cutting emissions worth using, including some that don’t look, at first glance, like an emissions lever at all.
This paper looks at one of those: how an organisation manages its computing infrastructure. Get that right, and it’s possible to cut direct energy use, support energy-saving ways of working, reduce operational costs, and meaningfully improve the organisation’s carbon footprint, all at the same time. Cloud computing is central to that shift.
Why this matters more than it used to
Data centres are no longer a minor line item in the world’s energy use. According to the International Energy Agency, data centres consumed around 415 terawatt-hours of electricity in 2024, about 1.5% of global electricity consumption, and that figure has been growing at roughly 12% a year since 2017.
The IEA’s central projection has that figure roughly doubling to around 945 to 950 terawatt-hours by 2030, driven largely by AI workloads, which are growing several times faster than conventional data centre demand. However this plays out, the direction is clear: computing infrastructure is becoming a bigger, not smaller, share of the energy conversation, which makes the efficiency of that infrastructure a bigger lever too.
The carbon impact of a given amount of computing also varies enormously by where it happens. Electricity generated from hydroelectric sources can produce a small fraction of the CO2 per kilowatt-hour of electricity generated from coal or gas, so moving workloads to fewer, larger, well-sited data centres, rather than running many smaller ones wherever convenient, is itself a genuine emissions reduction lever, independent of anything else.
What actually makes cloud more efficient
A large, well-run cloud data centre is not simply a bigger version of an on-premises server room. The efficiency gains come from a few specific, well-documented sources.
Operational efficiency
Commercial cloud services run more efficiently than smaller, on-premises deployments because of multi-tenancy and large-scale, dynamic provisioning. A large cloud provider can achieve utilisation rates far higher than a typical on-premises server room manages, which means organisations moving to the cloud typically end up provisioning a fraction of the servers they would have needed on-premises for the same workload.
Equipment efficiency
Cloud providers can adopt new, more energy-efficient hardware faster than most enterprises replace their own equipment, simply because of the scale they operate at. AWS’s move to its own Arm-based Graviton processors is a concrete example: Graviton-based instances use up to 60% less energy than comparable instances for the same performance, and more than 70,000 AWS customers have already used them.
Infrastructure efficiency
Advanced cooling, lighting, and facility design at large-scale cloud data centres reduce energy requirements that a typical on-premises data centre simply can’t match, since most organisations don’t operate at a scale where purpose-built, world-class facility design is economical.
Where the power actually comes from
Moving workloads to the cloud can shrink an individual organisation’s own footprint, but the combined footprint of the major cloud operators is enormous, which makes their choice of energy source the fourth, and arguably most consequential, lever.
AWS reports that 100% of the electricity it consumed in 2025 was matched with renewable energy, for the third year running, having reached that target seven years ahead of its original 2030 goal. Microsoft has now reached the same 100% match. Google reached its own 100% renewable match in the late 2010s and is now working towards running every grid it operates on carbon-free energy every hour of the day by 2030, reporting around 70% hourly matching in the US as of 2024.
None of this means a given kilowatt-hour drawn from the grid is necessarily carbon-free in the moment; it means these providers are purchasing enough renewable generation, overall, to match what they use, which is a meaningfully different claim worth understanding rather than taking at face value.
This is directly relevant to EnergySys itself. EnergySys runs on AWS, having held AWS Energy Competency status since 2021, which means every organisation running on EnergySys is, indirectly, running on infrastructure backed by AWS’s renewable energy commitments and its efficiency gains from custom silicon like Graviton.
AWS reports its infrastructure is up to 4.1 times more energy efficient than a typical on-premises data centre, and that moving a workload to AWS can cut its carbon footprint by up to 99% compared with running it on-premises. Choosing a platform’s underlying infrastructure is itself a small but real part of an organisation’s own path to net zero.
Electricity isn’t the only resource worth scrutinising either. Large data centres also use significant volumes of water for cooling, and that side of the environmental picture gets far less attention than carbon claims do. It’s worth asking a provider about water use and replenishment alongside energy sourcing, not just accepting a renewable energy percentage as the whole story. A genuinely sustainable provider should be able to speak to both.
Beyond electricity: what the data itself makes possible
The direct energy savings from moving to the cloud are real, but the bigger opportunity is what becomes possible once an organisation’s data is easy to manage and analyse together, rather than scattered across disconnected systems.
Smart building management is a useful, well-documented example outside oil and gas. Microsoft’s own Redmond campus, grown organically over decades into more than 500 acres and around 15 million square feet of office and lab space, became a test bed for smart building software running on Azure, pulling in data from thousands of sensors tracking heating, cooling, and lighting alongside external data like weather forecasts.
That combination caught faults that would previously have gone unnoticed. In one case, exhaust fans left running by mistake for a year wasted tens of thousands of dollars in energy before the anomaly was caught. The same principle, sensor data plus accessible analytics catching what a spreadsheet or a manual walk-through would miss, applies directly to production operations and infrastructure across the oil and gas value chain, not just to the buildings that house the people running it.
The same combination of cloud-based data and analytics has been applied well beyond conventional IT. Google’s collaboration with Stella McCartney used cloud-based data analytics to give the fashion brand visibility into a supply chain most brands can’t see past their immediate suppliers, in an industry responsible for a genuinely significant share of global wastewater and carbon emissions.
Separately, Global Fishing Watch, a partnership between Google, Oceana, and SkyTruth, built a free, public, interactive view of the world’s large industrial fishing fleet, tens of thousands of vessels, to support marine habitat protection and sustainable fisheries management. Neither of these is an oil and gas example, but both show the same pattern: once operational data is accessible and analysable at scale, insight and accountability follow that weren’t possible before.
The way people work has changed too
Cloud and mobile technology have made flexible, remote, and hybrid working the norm rather than the exception across most industries, including oil and gas, and that shift carries a genuine emissions benefit through reduced commuting and business travel.
It also extends further than the office: the same cloud and connected-device infrastructure that supports flexible office work also supports much higher levels of remote monitoring and management of physical infrastructure, globally, which matters even more for an industry whose assets are spread across offshore platforms, remote fields, and pipeline networks that used to require someone physically present to check on them.
Why this isn’t just a sustainability nicety
The energy an organisation’s IT infrastructure consumes doesn’t sit outside its formal emissions reporting. Purchased electricity, including whatever a cloud provider’s data centres draw from the grid on an organisation’s behalf, falls under Scope 2 emissions in both the GHG Protocol and Australia’s NGER scheme.
A company weighing up its own emissions numbers ahead of a Safeguard Mechanism baseline, or preparing for the assurance requirements phasing in under Australia’s climate disclosure regime, is already accounting for this whether it’s thought about IT infrastructure specifically or not. Choosing a more efficient, more renewably-powered platform is one of the more straightforward levers available, precisely because it doesn’t require redesigning a production process or waiting on a new technology to mature.
What to actually look for when choosing a provider
Moving to the cloud isn’t simply a decision to make once and forget. A few practical questions are worth asking of any provider, and revisiting periodically as commitments and capabilities change.
- What share of their electricity consumption is currently matched with renewable energy, and how has that changed over the last two or three years, not just what the target is.
- Whether that renewable claim is based on actual hourly or locational matching, or an annual, market-based offset that may not reflect what’s actually powering a given data centre at a given moment.
- What their water use and replenishment commitments look like, not just their carbon figures.
- What hardware efficiency gains they’re passing on, custom silicon, cooling design, utilisation rates, rather than just marketing language about scale.
- Whether moving a specific workload to their infrastructure is likely to reduce its footprint meaningfully, given the nature of that workload, rather than assuming any cloud migration automatically helps.
None of this is a case for treating IT infrastructure as a side issue in a net zero strategy. It’s one part of a much bigger set of changes energy companies need to make, but unlike many of those changes, it’s one an organisation can act on now, without waiting on a technology that doesn’t exist yet or a market that hasn’t matured.
Explore how EnergySys supports emissions and CCUS data management, or get matched with an EnergySys partner who can help you make the most of it.
