By Sally Bailey | 20 August 2026
Artificial intelligence is transforming how businesses operate, and understandably, questions have been raised about its environmental impact. It’s true that training and running AI models takes significant computing power, and that in turn increases energy consumption, water use and infrastructure demand, along with the emissions that come with it. Yet that’s only one part of the picture.
Across the technology industry, cloud providers, software vendors and technology firms are investing heavily in making AI technologies more efficient. In other words, the same technology driving increased demand is also being used to reduce waste, optimise resources and improve sustainability performance, and at a scale that wasn’t previously possible.
Of course the impact doesn’t stop at AI’s own infrastructure. Organisations across all sectors are increasingly using AI systems to support sustainability initiatives, from identifying energy conservation opportunities and optimising logistics to improving carbon emissions reporting and uncovering inefficiencies in resource use. Effectively, AI is helping businesses make more informed sustainability decisions at scale.
So, for business leaders, the conversation is shifting. The question is no longer whether AI consumes resources, because it does. The more important questions are whether AI can help organisations use those resources more efficiently, control the costs that come with them, and build a genuine competitive edge. And increasingly, the evidence suggests it can.
The infrastructure powering today’s AI platforms is a great example of this.
Much of AI’s energy demand runs through the cloud, so it makes sense that cloud providers are turning that same technology on the problem itself. Hyperscale providers such as Microsoft, Google Cloud and AWS are already using AI to optimise data centre cooling, predict workload demand, route processing to the most efficient resources available, and improve how large-scale infrastructure is designed and operated.
Location counts too. Many providers are now choosing to build new data centres closer to renewable energy sources such as wind and solar power, cutting carbon emissions by reducing reliance on carbon-intensive power generation from the outset. AI is playing a part here as well, with predictive tools helping providers forecast renewable energy availability and shift compute-intensive workloads to when and where clean power is most abundant.
This matters because hyperscale cloud data centres are typically far more energy efficient than traditional on-premise infrastructure. AWS, for example, reports that its cloud infrastructure is up to 4.1 times more energy efficient than on-premise environments, highlighting the scale of the efficiency gains that modern hyperscale platforms can deliver. And as AI systems become increasingly embedded in how these environments are managed and optimised, from workload allocation to cooling and energy forecasting, those efficiency gains are only expected to grow.
Energy isn’t the only resource under scrutiny though. Data centre cooling also relies on water, and this has become one of the most talked-about aspects of AI’s environmental impact.
The challenge is, many of the statistics circulating online are poorly sourced, inconsistent or presented without context. That doesn’t mean water use isn’t an important issue, of course, but it does mean businesses should look beyond viral headlines and focus on verified information.
The good news is that the industry is already investing in research and innovating rapidly. Microsoft, for example, highlights how cooling technology is evolving, noting that its latest AI-ready data centre design uses zero water for cooling during normal operations, saving more than 125 million litres of water per facility each year compared with traditional approaches.
While there’s a growing focus on the sustainability of these hyperscale data centres, it’s worth looking closer to home too. AI relies on a complex chain of infrastructure, and the network is a critical part of that chain. Get it wrong, and an inefficient network can create performance bottlenecks, drive unnecessary hardware investment and make it harder to scale AI services effectively.
That’s why many organisations are now reviewing their network alongside their cloud and AI strategy. Because modern, intelligently managed networks improve performance, enhance visibility and support more efficient use of resources across the estate.
And when AI, cloud and network infrastructure are designed to work together, organisations are far better placed to deliver both business outcomes and sustainability goals.
Infrastructure is only part of the story, though. How organisations consume technology matters just as much.
The cloud’s consumption-based model already helps businesses avoid overprovisioning resources, and AI is making that process even more effective. AI-powered monitoring tools can analyse usage patterns in real time, helping organisations match capacity to actual demand rather than relying on estimates. The result? Less idle infrastructure, lower operational costs and reduced waste.
Reducing on-premise infrastructure brings further benefits too, lowering maintenance requirements, reducing hardware refresh cycles and improving overall energy efficiency. AI-driven optimisation helps organisations get more value from the resources they already have before investing in additional capacity. In practice, these gains make efficiency a direct driver of cost savings, not just an environmental consideration.
Perhaps the biggest sustainability opportunity of all, though, comes from data.
Before organisations can use AI effectively, they need to understand, organise and govern the information that powers it. Poor-quality, duplicated or fragmented data doesn’t just create reporting challenges. It can also make AI less effective, requiring more processing, more intervention and ultimately delivering less value.
That’s why many organisations have found that successful AI adoption really starts long before the first AI use case is deployed. Strong data foundations help AI work more efficiently, improve decision-making and provide better visibility, insight and accountability into resource usage, emissions and operational performance.
And as sustainability reporting expectations continue to evolve, organisations with accurate, well-governed data will be better positioned to meet compliance requirements, spot opportunities for improvement and scale AI with confidence.
Not every AI application deserves the energy, water and compute it consumes. Before organisations optimise infrastructure, right-size consumption or strengthen data foundations, there’s a more fundamental question worth asking: is this the right use case for AI at all?
Every AI-generated output, however small, draws on real-world resources. That’s easy to forget when novelty applications go viral, the kind of quick, low-value content generation that spreads widely online but delivers little to no business benefit. These use cases might seem harmless, but at scale they carry real risks and represent a genuine, avoidable draw on natural resources, for a return that rarely justifies the cost.
For business leaders, the lesson isn’t to avoid AI. It’s to be deliberate about where it’s applied. The organisations getting the most from AI aren’t using it because everyone else is; they’re using it where it solves a real problem, be that identifying energy conservation opportunities, optimising logistics, improving emissions reporting or uncovering inefficiencies in resource use, as outlined above. That distinction between AI used for genuine value and AI used for novelty is quickly becoming a marker of maturity, and it belongs at the top of any responsible AI strategy, not an afterthought.
Put simply, the most sustainable AI decision an organisation can make isn’t whether to use AI, but where and why.
AI is moving from experimentation to everyday business use, and this shift is creating a clear divide between organisations that treat sustainability as a constraint and those that view it as a source of competitive edge.
As adoption accelerates, businesses are under growing pressure to control costs, improve efficiency and meet sustainability commitments, often all at the same time. That creates a real test for leadership teams, because AI initiatives are expected to deliver innovation and productivity gains, but they need to do so responsibly and cost-effectively, with clear accountability.
The organisations gaining competitive advantage aren’t the ones avoiding AI because of environmental concerns. They’re the ones deploying it strategically, efficiently and with the right governance in place, turning what could be a cost and compliance challenge into a source of differentiation instead.
Approached in the right way, sustainability and AI become aligned objectives rather than competing priorities, and that alignment is increasingly what separates leaders from the rest of the market.
Validate the use case before you scale. Before adopting or expanding any AI innovation, ask whether it solves a genuine business problem and delivers measurable value, not just novelty or convenience. This keeps AI adoption purposeful and avoids unnecessary resource consumption from low-value use cases.
Strengthen your data foundations. Start by assessing the quality, structure and governance of your data before scaling AI further. Well-organised, accurate data helps AI models work more efficiently, supports better sustainability reporting, and reduces the processing and waste that comes from poor-quality information.
Start with an audit, then right-size to demand. Begin by establishing a clear baseline of your current energy, cost and utilisation profile, then use AI-powered monitoring to match capacity to actual demand rather than relying on estimates. Right-sizing in this way often delivers some of the fastest cost and efficiency improvements available.
Review your network as part of your AI strategy. An inefficient network can create performance bottlenecks and drive unnecessary hardware investment as AI scales, so it’s worth reviewing your network alongside your cloud and AI strategy. Doing so improves visibility and supports more efficient use of resources across the estate.
It’s clear the sustainability debate around AI is becoming more nuanced. AI-optimised infrastructure is improving data centre efficiency, new cooling technologies are reducing water consumption, intelligent networks and cloud optimisation are helping organisations do more with less, and a sharper focus on choosing the right use cases is making sure every AI deployment earns its resource cost.
None of this means the challenge is solved, of course. But it does mean real progress is being made, and the conversation is changing.
The organisations seeing the greatest benefit aren’t treating sustainability and AI as competing priorities. Instead, they’re recognising that the right technology foundations, and the right use cases, can support both, and that is increasingly where long-term competitive advantage will be won.
Looking to scale AI without scaling waste? Our complimentary Data and AI Readiness Assessment identifies inefficiencies across your data estate and gives you the insight to build the foundations for sustainable, scalable AI adoption.
Book your Data & AI Readiness Assessment
26 August 2026 | By James Thomas
Discover four key challenges affecting AI adoption and how organisations can build confidence in their approach.
Read more
6 August 2026 | By Chris Thomas
Discover why more IT leaders are reviewing their business connectivity strategy and whether a 10Gbps bearer is the right next step.
Read more
24 July 2026 | By Jesse Brett
AI tokens are central to AI cost and consumption. Learn how to manage usage, control spend and scale AI with confidence.
Read more