Top Trumps: Sustainable IT Edition
- 3 hours ago
- 5 min read
Sustainable habits are already part of our everyday routines. We turn off lights when leaving a room, bring reusable bags to the supermarket and sort our recycling with the quiet confidence of someone who definitely knows which packaging goes in which bin. These small actions show how sustainability has quietly become embedded in how we live.
But step into the world of computing and everything changes. It’s no longer a case of remembering our reusable water bottle, being green means navigating massive volumes of code, complex algorithms, and workflows that can require vast levels of energy consumption. As of November 2021, the Association for Computing Machinery estimated that Computing and IT account for between 1.8% and 3.9% of global greenhouse gas emissions. [1] This is only increasing and, based on current trends, could exceed a third of global emissions by 2050.
Sustainability in development and technology has never been more relevant. And yet, it is still widely treated as something handled by ‘the organisation’, or ‘the cloud team’, or ‘the sustainability working group we think meets on Wednesdays’. In reality, those who can have an immediate influence on sustainability are often individual analysts, developers and engineers - those who write the queries, design the models, schedule the jobs, and spin up (and occasionally forget to spin down) the environments.
As a consultancy firm with a focus on cloud computing, artificial intelligence and wider emerging technologies, Principle One has a responsibility to understand the landscape we work in and make practical and sustainable choices wherever we can. To keep this front of mind in our day-to-day work, we've developed a set of Sustainable IT Top Trumps. Each card highlights a practical action or decision that can reduce environmental impact. Some focus on small, everyday habits, like turning off unused environments or writing efficient queries, while others emphasise optimising workflows and pipelines for better efficiency. The cards are designed to make it easy to see where your choices matter most, and to show that sustainable IT isn’t something abstract, it’s something you can influence every time you sit down at your computer. In this article, we’ve focused in on just a few of them to show how small engineering lifestyle choices add up.
Efficiency starts in the WHERE clause
Efficient query design is one of the most immediate ways analysts can reduce the environmental footprint of their work. Databases, after all, are incredibly obedient. They will execute exactly what you ask, no matter how over ambitious or wildly inefficient your SQL may be. Tightening up queries by selecting only the fields you genuinely need, avoiding unnecessary joins, and filtering data early on all help reduce the amount of processing (and therefore energy) required to return results. Indexing and avoiding full table scans where possible can make a big difference too. And while it’s tempting to run expansive queries ‘just in case’, designing them deliberately keeps systems faster, costs lower, and carbon usage noticeably lighter. The impact scales quickly, especially in teams that run similar workloads day after day.
Not all languages cost the same
The choice of coding language is often driven by a number of factors: the skills the team already have, the skills they would like to develop, or whatever happens to be popular in the job market that month. Very rarely is the decision influenced by sustainability, which is surprising given that it can make a huge difference. Take Python for example. It’s one of the most popular languages globally, but it’s also one of the most energy-hungry languages around, consuming roughly 75 times the energy of C[2]. That’s the equivalent of running your washing machine for one sock. Thankfully, advances in architecture such as containerisation and micro services mean that you no longer need to take an ‘all or nothing’ approach. Instead, you can progressively shift individual components to lighter, more efficient frameworks without affecting the whole system. Bringing AI into the mix, there are now engines that can be used (with caution) to migrate an energy hungry service running Python to something much more efficient like C or Java.
Smarter prompts, smaller footprint
Prompt engineering is another area where you can directly influence energy consumption, which is increasingly important as AI models become more embedded in everyday systems and wider working practices. Large language models (LLMs) can be extremely power hungry, so crafting prompts that are clear, concise and well-structured helps keep the computational effort (and electricity bill) under control. Providing the right context upfront can prevent multiple follow up queries, while using more targeted prompts can reduce the size or complexity of the model needed in the first place. Even small changes, like asking for shorter outputs or reusing cached responses, cut down on unnecessary processing cycles. While these adjustments might seem minor, they add up quickly, making good prompt engineering a meaningful way to reduce the carbon footprint of AI‑powered systems.
Let the grid breathe
Time based scheduling can also be optimised to run workloads that are not time critical during periods where there is surplus renewable energy available through the grid. Just as most EV owners charge their vehicles overnight to utilise surplus wind energy during otherwise quiet periods, it is possible schedule IT systems to run their processes in a way that will minimise their carbon footprint. In some regions where the cost savings are passed on to consumers, using surplus energy outside of peak hours can also result in money savings (win-win). Examples of workloads that may be able to benefit from this technique include: data migration activities, data reconciliation tasks, bulk operations, refreshing large datasets, and updating standalone AI models.
Every decision counts
Whether decisions are strategic or hands‑on, they all contribute to a more sustainable approach to technology. Regardless of your role, each and every one of us working in the technology sector should have sustainability built in to how we work. The choices may seem small at times, from selecting a more efficient language to remembering to spin down that dev environment that’s been abandoned all week, but they add up quickly when they’re repeated across teams and organisations. At Principle One, we’re conscious there’s still a way to go to make our development as green as it can be. But every day we continue to educate ourselves and prioritise the most impactful changes we can make to contribute to a more environmentally friendly future of consultancy.
Comments