As New Zealand races to capitalise on surging global demand for AI infrastructure, a critical question sits beneath every major data centre announcement: can the grid actually support it? According to Sam Kivi, Chief Technology Officer at Grid Share, the answer depends less on how much capacity exists and more on how it is used.
“A lot of work has gone into improving the process for grid connections,” says Kivi, “however large projects can still take several years to achieve connections.”
Location, projected demand profile and infrastructure alignment all factor into that complexity. And in a market moving at pace, Kivi is direct:
“We have seen proposed locations for major data centre projects in recent times that are wholly inappropriate from an energy resource and infrastructure perspective.”
The grid connection challenge is real. But Kivi’s view is that it is also, in part, a problem of imagination.
The assumptions that leave capacity on the table
The most common misunderstanding Kivi encounters is the assumption that data centre demand is inherently inflexible. That assumption, he argues, leads to conservative connection solutions that underutilise available resources and unnecessarily extend timelines.
The traditional approach prioritises certainty over optimisation and while that can simplify approval processes, it is not the only path forward.
The opportunity, as Kivi sees it, is for project developers to lead differently. Bringing flexibility and demand response solutions to the table early, rather than treating them as an afterthought. It reframes the developer’s role from passive applicant to active partner in grid management.
What an energy-responsive data centre actually looks like
For many in the industry, energy flexibility remains an abstract concept. Kivi makes it concrete.
Grid Share’s model sits at one end of what he describes as a broad spectrum of responsiveness, a facility designed with energy flexibility as a core priority, not a secondary feature. In practice, that means a facility capable of curtailing 100% of its load in milliseconds for grid stability services, in seconds to minutes for energy arbitrage and hedging, and for months at a time for seasonal renewable resource utilisation.
“These are highly modular, relocatable, lightweight plug-and-play solutions,” says Kivi, “but they serve a niche computing market.” The more significant opportunity, he believes, lies in translating these principles into more conventional data centre facilities. Incorporating flexibility features that ease grid connection challenges, improve renewables utilisation and reduce reliance on the carbon-emitting components of the electricity system.
The spectrum exists. The question is how much of it developers are willing to engage with.
What developers entering the market should prioritise
For those navigating grid strategy now, Kivi’s advice is to start with the customer and think carefully about where computing demand is heading.
“Look closely at their customer base now and into the future and consider the growth of flexibility as an enabler of greater grid-data centre co-optimisation,” he says. Compute tasks are increasingly stratifying into varying layers of latency dependency and time-criticality. Understanding that stratification and taking it into conversations with grid operators and energy providers early changes what is possible.
“A little flexibility procured up front can be the signal for faster and larger connections and lower overall emissions”
says Kivi. It is a counterintuitive insight: offering to do less, under certain conditions, can unlock the ability to do significantly more.
The relationship that needs to change
Looking five years ahead, Kivi’s vision is of a fundamentally different dynamic between data centres and the electricity grid, one built on real-time co-ordination rather than static agreements.
“I would like to see an established market for demand response services that align with the price and carbon emission points that various forms of data centre customers target,” he says. The result would be a live, dynamic management of data centre load in co-ordination with grid operators, generating clear demand signals, driving greater utilisation of renewable energy resources and reducing the overall cost of electricity supply for all New Zealanders.
The potential, as Kivi frames it, is not incremental.
“There’s huge potential for more give and take on both sides of the relationship and a better outcome for New Zealanders and the New Zealand economy.”
A grid and a data centre sector that operate in genuine partnership, rather than in parallel, represents a material shift in how New Zealand manages its energy future.
For an industry accustomed to treating the grid as a constraint to be navigated, that reframe may be the most important infrastructure decision of all.
Grid, flexibility and the future at Data Centre Leaders Summit New Zealand
Sam Kivi’s perspective cuts to one of the most consequential and least visible challenges facing New Zealand’s data centre ambitions. At Data Centre Leaders Summit New Zealand, where power, planning and capital align to build New Zealand’s digital future, his session brings the grid into focus: not as a barrier, but as a dynamic system that the industry has both the opportunity and the responsibility to shape.
For developers, investors and infrastructure leaders planning for the next decade, understanding the grid flexibility frontier is no longer optional. It is the difference between projects that connect and projects that stall.
