EFFICIENCYIT
Future challenges in an AI-driven era Nick says the industry now faces unprecedented change.
“ We are headed into uncharted territory, genuinely, because our world has literally been turned on its head by Nvidia,” he says.
The rapid escalation of rack power density, compute demand and cooling complexity is redefining design approaches.
Yet he views this as a phase of opportunity for those adapting early.“ AI and this new world of tech is allowing us to pivot how we use the data centre. We’ re now working on a project where we have around 5MW of 55-degree water as the output from a supercomputing cluster. Our plan is to give all of that back to a nearby district heating system.”
That scheme, he says, could prevent nearly a million litres of kerosene being used for heating.“ That is the kind of technology and the sustainability use cases that we are working towards.”
Nick concludes that decentralised models will increase as the UK looks to use waste heat more efficiently.
“ The requirement to have a data centre in one of the UK availability zones has been reduced, so we can decentralise. And that means we can put our data centres into places where we can use the waste output much more efficiently.”
Here,‘ efficiency’ is the operative word – the clue is in the name, after all. sustainabilitymag. com 91