The green data centre boom: how New Zealand could be at the forefront

By Jason Molloy, Country Leader, Secure Power, Schneider Electric

Generative Al is taking off even faster than the smartphone or the Internet. Having first reached public attention with the launch of ChatGPT in late 2022, it's now being used in nearly every large organisation across New Zealand, according to a recent Microsoft report. That spells significant extra demand for data centres to store and process the huge amount of data genAl requires. McKinsey suggests that global demand for data centre capacity could more than triple by 2030, spurred by Al's phenomenal growth.

In order to leverage genAl to its full extent, Kiwi organisations need to consider where they store their data. But is bigger always better when it comes to data centre infrastructure? Not necessarily. The truth is that businesses need to consider the data centre model that works best for their circumstances. For more and more organisations, that's a hybrid cloud solution which offers the best of both worlds, enabling them to share the innovation that comes with access to public cloud services while maintaining critical data close to home.

At first glance, it seems logical to go all-in on public cloud. Data centres owned by the big global cloud providers typically feature world-class security thanks to the amount they can invest, as well as being specially equipped to host advanced Al workloads at enormous scale. In New Zealand, new data centres are being opened by a host of cloud providers, from multinationals Canberra Data Centres (CDC), AWS and Microsoft to local companies Datacom, Spark, Catalyst Cloud and others, all promising to help customers get the most out of the latest technologies.

Having access to off-site public cloud storage also provides greater resilience should anything happen to on- premises servers - which, given New Zealand's susceptibility to natural disasters such as earthquakes and floods, is a risk well worth considering. The pay-as-you-go consumption model of public cloud services is also very attractive compared with the costs of maintaining your own infrastructure, particularly when demand is lower and you're carrying the costs of equipment you don't need.

However, there are also cost benefits to maintaining on-premises edge sites. In a region-wide survey by Schneider Electric, 84% of respondents saw cost reductions of up to 25% through adopting edge computing (distributed computing that moves data processing and storage closer to the source of the data, rather than relying solely on centralised data centres), reducing the bandwidth required to transfer huge amounts of data.

Also, edge infrastructure may still have years of life left. Especially at a time when businesses are being challenged by tough economic circumstances, it may make more sense to upgrade existing systems rather than abandon significant capital investments. Public cloud migration is not always a simple lift and shift. It often requires re-engineering.

The result is, organisations move systems gradually, as and when it makes sense for their business, running multiple IT estates across on-premises and public cloud. Datacom's 2024 Annual Cloud Report showed a notable shift towards a hybrid cloud strategy among New Zealand organisations, where workloads are managed on a case-by-case basis, seeking to balance performance, efficiency and cost-effectiveness.

The report also showed a significant focus on private cloud to support Al solutions, for privacy reasons, backing up Schneider's own data showing nearly 40% of Kiwi financial services businesses expect to use their own data centres for Al- related applications. Many organisations have concerns about enabling large language models to be trained on their own personal data or IP, while embracing public cloud 's scalability and automated security for other business functions.

The hybrid approach also makes sense where latency is concerned. Until now, this has been a major challenge for many operations, such as investment brokers or airports, whose decision-making requires data to be accurate within fractions of a second. That's traditionally meant such organisations prefer private cloud infrastructure, but this can also hinder growth. With hyperscalers (large cloud service providers) now arriving in New Zealand, public cloud latency is significantly reduced, but businesses may still want to keep their own data close to maintain sovereignty.

The good news is that it doesn't have to be an either/or choice between going public and owning your own micro data centre. These organisations now have the choice to retain certain workloads in edge infrastructure (including easily-upgraded mobile data centres for maximum cost-effectiveness), while leveraging the scale of public cloud for "spillover” at times of extremely high demand.

Designing local data centres with smart AI powered tools that monitor and minimise consumption can help reduce the environmental burden for all organisations, regardless of their size and budget. Many organisations with private data centres have struggled to ensure energy use is optimised throughout the infrastructure's life, as it ages. At the same time, even using the most sustainable public cloud in the world, Kiwi businesses still need to ensure taking advantage of its unlimited scale isn't blowing out consumption and costs unnecessarily. AI-enhanced tools that manage cloud and energy use across their entire IT environment make hybrid models much simpler, while giving them the best of both worlds.

In an Al- driven world, organisations don't need to choose between global scale or data sovereignty. Thanks especially to new Al tools, a hybrid data centre model has become the smartest option.