AI SUSTAINABILITY
Sustainable solution? AI is becoming increasingly used by many businesses, but it is power-hungry so how do businesses square that with their ESG goals – and how can resellers help with this?
The rapid adoption of AI by many businesses across the spectrum to carry out a range of tasks and functions is anticipated to continue this year. While AI can bring benefits to those that use it, AI can also bring challenges – not least to sustainability targets. Jason Beckett, head of technical sales EMEA at Hitachi Vantara, says the demands of AI are rapidly outpacing traditional IT infrastructure. “AI workloads, particularly those running on GPU- intensive platforms or training large language models, require substantial computing power,” he adds. “This often translates into higher energy usage, increased cooling needs, and a greater carbon footprint especially when deployed without optimising for energy efficiency. “These pressures are already under scrutiny from regulators, investors and sustainability-conscious customers. This focus will only intensify. Many legacy systems weren’t built to handle AI at scale, particularly the associated cooling demands. “If left unaddressed, this creates real risk of grid overload and unsustainable operational costs. To stay ahead of these challenges, businesses must rethink their infrastructure strategies now. Innovation depends on agility and scalability: if your infrastructure can’t adapt quickly, every new initiative becomes a bottleneck. By designing for simplicity, security, and sustainability, organisations can pivot to new technologies like AI without major rework.” For many businesses, the challenge isn’t just the initial training of models, but
With AI use increasing, and the carbon produced by it falling under scope 3 emissions, this poses a potential problem for businesses that want to use AI but are also mindful of their sustainability targets.
the ongoing cost of inference – systems running continuously in the background, adds Adam Herbert, Go Live data CEO. “What’s often underestimated is how quickly these demands scale when AI is deployed without clear constraints or purpose,” he says. Richard Eglon, CMO Nebula Global Services, agrees that AI is inherently power- hungry. “This creates an ironic challenge: businesses are using energy-intensive technology to meet environmental goals,” he says. “Data centres, for example, need to manage cooling systems and server configurations carefully because every operational change impacts energy use and carbon output.” Simon Evans, lead consultant and sustainability director at PIE Factory, notes that the carbon produced by processing AI data requests falls under scope 3
22
Powered by FlippingBook