New Zealand is taking a significant step into the global AI infrastructure race.

Hyperscaler Datagrid has received full resource consent to build a 280MW data centre in Southland, a facility that could become the country’s second-largest electricity user once operational. The approvals were granted by Southland District Council, Environment Southland, and Invercargill City Council, enabling the development of what the company is describing as New Zealand’s first “AI factory”.

The project is being led by Rémi Galasso, known for founding the Hawaiki submarine cable, which connects Auckland with the United States and significantly improved New Zealand’s international connectivity.

This development places New Zealand more firmly in the rapidly expanding global market for AI infrastructure and hyperscale data centres.

What an “AI Factory” Actually Means

The term “AI factory” reflects a shift in how data centres are being designed.

Traditional facilities primarily store and process enterprise data or support cloud workloads. AI infrastructure is different. Training and running large AI models requires enormous computational power, specialised hardware, and high-density processing environments.

These facilities typically include:

• Large clusters of GPU-based compute infrastructure
• High-capacity networking fabrics
• Advanced cooling systems
• Massive energy consumption

The result is infrastructure designed not just for storage or applications, but specifically for AI model training and inference.

Globally, this shift is accelerating rapidly.

The Global Expansion of AI Data Centres

Hyperscale data centres are expanding worldwide as demand for AI computing capacity continues to grow.

Major technology companies are investing heavily in AI infrastructure, with new facilities being announced across North America, Europe, and Asia. AI workloads require significantly more power than traditional cloud computing environments, often leading to facilities measured in hundreds of megawatts.

To put the Southland proposal in context:

• 280MW places the facility firmly in hyperscale territory
• Large AI data centres globally now range between 200MW and 500MW
• Power availability has become the primary constraint for AI infrastructure growth

This explains why many projects are being developed in regions with access to reliable and renewable electricity.

Why Southland is Attractive for Hyperscale Infrastructure

Southland offers several characteristics that make it attractive for a large data centre project.

These include:

• Access to significant renewable energy generation
• Cooler climate conditions supporting data centre cooling
• Availability of land suitable for large infrastructure projects
• Existing international connectivity through submarine cable infrastructure

These factors are increasingly important as hyperscale operators look beyond traditional technology hubs to support growing AI compute demand.

New Zealand’s renewable electricity mix also adds an additional layer of appeal for companies seeking lower-carbon AI infrastructure.

What This Means For New Zealand’s Digital Economy

The project highlights a broader shift in how countries compete in the global digital economy.

Historically, data centres followed population density and enterprise demand. AI infrastructure is beginning to follow energy availability and connectivity instead.

For New Zealand, developments like this could:

• Attract international technology investment
• Strengthen the country’s role in global AI infrastructure
• Support growth in digital services and research
• Increase demand for skilled engineering and cyber security capabilities

At the same time, large infrastructure projects inevitably raise questions around energy allocation, environmental impact, and long-term economic benefit.

These debates are already happening globally.

The Cyber Security Dimension

Large-scale AI data centres also introduce important cyber security considerations.

These facilities often support:

• Sensitive model training data
• High-value intellectual property
• Critical digital services and platforms

This means the security architecture surrounding hyperscale facilities must address:

• Application security and AI workloads
• Supply-chain and hardware security
• Identity and access management at scale
• Data sovereignty and regulatory requirements

As AI infrastructure grows, AI in cyber security is also becoming increasingly important to detect threats across large, complex environments.

Lessons From the UK and Europe

The UK and European markets have seen similar developments over the past few years as AI infrastructure expands.

Key lessons emerging from those markets include:

• Energy capacity often becomes the biggest constraint
• Planning approvals increasingly factor in environmental impact
• National governments are recognising data centres as strategic infrastructure
• Cyber resilience expectations rise as infrastructure becomes more critical

New Zealand is now entering a similar phase of infrastructure growth.

A Moment of Opportunity

The Southland project signals something broader than a single facility.

It reflects the increasing importance of AI infrastructure, hyperscale data centres, and national capability in the global technology landscape.

For New Zealand, the challenge will be balancing innovation, infrastructure growth, energy demand, and digital resilience.

If you are planning for infrastructure, connectivity or security implications of AI and hyperscale environments, we can help you assess how your network and cyber security strategy aligns with the next generation of digital infrastructure, to find out more contact us below….

📞 UK +44 (0) 113 341 0123

📞 NZ +64 (0)9 802 2444

📧 hello@itogether.com