Hosts supply the land and energy. Decisions happen elsewhere.
As artificial intelligence reshapes the global economy, New Zealand finds itself courted by some of the world's largest technology companies, drawn by its renewable energy, stable climate, and open land. Datagrid's planned AI facility near Invercargill and AWS's NZ$7.5 billion Auckland investment promise jobs and connectivity, yet the deeper pattern is an old one: smaller nations supply the ground, the power, and the infrastructure, while the decisions — and the highest-value returns — remain in the hands of those who own the platforms. The question New Zealand must sit with is not whether to participate in this global build-out, but on whose terms, and toward whose long-term benefit.
- A global race to build AI infrastructure has reached New Zealand's shores, with billions of dollars and enormous electricity demands arriving faster than policy frameworks can absorb them.
- Datagrid's Invercargill facility alone could consume up to 280 megawatts — roughly 6 percent of the nation's total electricity — placing extraordinary pressure on a grid built for a small country.
- Local technology firms are not being displaced so much as repositioned: pushed toward land acquisition, power negotiation, and facility construction while foreign giants retain control of the software and platforms that generate the most value.
- Government agencies have actively marketed New Zealand as an ideal host nation, but the seductive logic of investment and jobs risks obscuring a structural imbalance in who captures lasting economic and strategic benefit.
- The path forward demands that policymakers distinguish between the appearance of participation in the AI economy and the substance of it — asking not just what is being built here, but what is being decided here.
New Zealand is caught up in a worldwide scramble to build the physical infrastructure that artificial intelligence depends on. Singapore-based Datagrid is planning what it calls the country's first AI factory near Invercargill — a 78,000 square metre facility costing billions — while Amazon Web Services announced a NZ$7.5 billion investment in Auckland data centres last year. Officials have welcomed these projects, pointing to New Zealand's renewable energy, cool climate, and political stability as competitive advantages, with some suggesting the country could become an international data centre hub.
The surface appeal is real: jobs, investment, and improved digital connectivity. But the distribution of value is far less even than the headlines imply. Datagrid's facility is expected to draw up to 280 megawatts of power — roughly 6 percent of national electricity consumption — making it the country's second-largest electricity user. That appetite reflects the computational scale required to run the AI systems powering tools like ChatGPT, but the electricity demand is only one dimension of the story.
Research on global data centre markets reveals a recurring pattern in smaller economies: infrastructure is built locally, but the systems it supports are controlled from elsewhere. Datagrid has positioned New Zealand as a host for international AI workloads dominated by Amazon, Microsoft, and Google. Local firms, unable to match the capital of global players, increasingly find themselves in infrastructure roles — securing land, negotiating power access, building to overseas specifications — while foreign companies provide the platforms and software running on top.
For a government thinking beyond the immediate economic cycle, this raises uncomfortable questions. If domestic firms become concentrated in lower-value infrastructure work, what does that mean for building higher-value capability over time? The benefits of these investments are genuine, but their lasting impact will depend on how carefully New Zealand manages the trade-off between participation and sovereignty in the emerging AI economy.
New Zealand is in the middle of a global scramble. Across the world, tech companies are racing to build the physical backbone that artificial intelligence runs on—the vast data centres that consume enormous amounts of electricity and computing power. That race has arrived here, and it's moving fast.
Singapore-based Datagrid is leading the charge, planning to open what it calls New Zealand's first "AI factory" near Invercargill. The scale is hard to ignore: a facility spanning 78,000 square metres, costing billions of dollars. It won't be alone. Amazon Web Services announced a NZ$7.5 billion investment in a cluster of data centres in Auckland just last year. On the surface, these projects look like straightforward wins for a small country—jobs, investment, a seat at the table of the global AI economy.
But the story underneath is more complicated. New Zealand's government and economic development agencies have been actively courting these investments, pointing to the country's renewable energy, cool climate, available land, and political stability as reasons foreign investors should build here. Officials have even suggested New Zealand could become an "international data centre hub." Datagrid itself has marketed the Invercargill project as "the most significant upgrade to New Zealand's digital infrastructure in a generation." The pitch is seductive because it contains real truth—these projects will bring money and employment. Yet the distribution of that value, and who ultimately controls it, is far less even than the headlines suggest.
The electricity demand alone tells part of the story. Datagrid's facility is expected to draw up to 280 megawatts of power—roughly 6 percent of New Zealand's total national electricity consumption. That would make it the country's second-largest electricity user, behind only the Tiwai Point aluminium smelter. These systems are power-hungry because they run the kind of artificial intelligence that powers ChatGPT and Claude, systems that require vast computational resources. But the deeper issue is not the electricity itself. It's where the value flows.
Research on global data centre markets reveals a pattern that repeats across smaller economies: the infrastructure gets built locally, but the systems it supports operate across multiple countries and are controlled from elsewhere. Datagrid has made clear it expects to serve international AI and cloud providers, positioning New Zealand as a host for global AI workloads. Those workloads come from markets already dominated by a handful of major tech companies—Amazon, Microsoft, Google. This concentration raises a fundamental question: where is value actually created, and how much of it stays in New Zealand?
The answer lies in how the work gets divided. Large international investors have the capital and market reach to build and operate these facilities, something smaller domestic cloud service providers simply cannot match. So what happens to those local companies? They don't disappear. Instead, they shift. As AI infrastructure demand grows, local firms increasingly take on the physical, unglamorous side of the business: securing land, negotiating access to power supplies, building facilities to overseas specifications, navigating local and international regulations. Meanwhile, the global tech giants provide the platforms and software that run on top—often by leasing capacity from those local operators. Amazon Web Services, for instance, recently moved away from building a large standalone data centre in Auckland, but it continues to expand its cloud operations through agreements with local data centre providers. The arrangement works, but it's uneven. Hosts supply the land, energy, and networks. Decisions about how systems operate are made elsewhere.
For a government thinking long-term, this raises hard questions. If domestic firms end up concentrated in infrastructure roles, what does that mean for developing higher-value capability over time? How should countries weigh immediate gains—jobs, investment, better connectivity—against longer-term concerns about control and positioning in the digital economy? These projects will deliver real benefits. But their true impact will depend on how trade-offs are managed and how much influence New Zealand retains over the systems it hosts.
Citas Notables
The most significant upgrade to New Zealand's digital infrastructure in a generation— Datagrid, describing its Invercargill project
While the infrastructure is built locally, the systems it supports usually operate across multiple countries— Angus Dowell, University of Auckland researcher on data centre markets
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter where the value goes? Isn't a job in Invercargill still a job?
It is. But there's a difference between a job and an industry. If you're always building the infrastructure for someone else's platform, you never develop the expertise to build your own platforms. You stay dependent.
So you're saying New Zealand could end up as just a landlord?
Not just a landlord—a landlord with very specific constraints. You're building to overseas specifications, following their rules, hosting their systems. The strategic decisions happen in San Francisco or Singapore.
But AWS is investing 7.5 billion dollars. That's real money.
It is. And it will create jobs and tax revenue. But most of that value—the profit from the software, the control over how the systems work—flows back to the company. New Zealand gets the electricity bill and the employment.
Is there a way to do this differently?
Potentially. If New Zealand required local partnerships, or invested in building its own high-value capabilities alongside the infrastructure, it could capture more of the value chain. But that requires governments to think beyond the immediate investment announcement.
What happens if they don't?
You end up with world-class infrastructure that serves global interests first. That's not necessarily bad, but it means New Zealand's long-term positioning in the digital economy is shaped by decisions made elsewhere.