July 31, 2026

Decades of routing through Auckland are over as Invercargill gets direct cable to Melbourne

USS Frank Cable

A digital dead end, finally getting an exit

The South Island has always been a branch office of New Zealand’s internet. Every international connection, from a Dunedin research lab to a Queenstown tourism operator, has to travel north through the North Island before it reaches a submarine cable to the rest of the world. That detour adds latency and, more importantly, concentrates the entire risk of an outage on a single set of northern routes.

That is about to change. Singapore-owned Datagrid NZ has signed contracts with Alcatel Submarine Networks and OMS Group for the supply and installation of a 2,300km submarine cable running directly from Invercargill to Melbourne. It is the first international submarine cable ever to land in the South Island, and it forms the first leg of Datagrid’s Tasman Ring Network, a roughly 6,000km loop that will eventually link both islands with Sydney and Melbourne.

The number that makes this concrete for business is latency. Datagrid chief executive Rémi Galasso says the cable will improve latency by up to 45 percent compared with existing routes via the North Island, calling it a “game changer” that would “unlock the potential of the region, and will allow South Island IT companies, universities and research centres to benefit from better internet.”

The data centre is what pays for the cable

Submarine cables do not get built on goodwill. They get built when someone needs the bandwidth badly enough to underwrite the economics. Here, that anchor is Datagrid’s proposed 280MW hyperscale data centre at Makarewa, 7km north of Invercargill, expected to cost between NZ$3.3 billion and NZ$3.5 billion. It already holds resource consents.

The scale is worth pausing on. Once running, the facility would be New Zealand’s second-largest electricity user, behind only the Tiwai Point aluminium smelter. That is the point. The data centre generates the traffic that justifies the cable, and the cable exports the data centre’s output to overseas customers. Neither works alone.

The most credible outside validation the project has received came the week before the cable announcement, when Mercury took a 12.7 percent stake in Datagrid NZ for NZ$53 million. When a listed generator writes that kind of cheque, it is a signal the project is past the PowerPoint stage.

Years of surveys, now real contracts

This has been a long time coming. The project ran under the Intelia NZ name before rebranding to Datagrid, and route surveys across the Tasman began in 2024 with EGS Survey to find the shortest, most secure path. Earlier documentation for the system, then branded Te Waipounamu, described 120Tbps of capacity across four fibre pairs, though those specs predate the current Tasman Ring framing and should be treated as indicative.

Signing supply and installation contracts is the material step beyond all of that. It is the difference between a feasibility study and a project with steel and fibre attached to it.

The demand curve keeps climbing

The timing lines up with a bandwidth-hungry country. Chorus reported in March 2026 that national fibre connections had reached 1.13 million, uptake had hit 72.4 percent of serviceable addresses, and monthly average data usage per connection climbed to 722 GB in January 2026. Chorus chief executive Mark Aue said the company was targeting 80 percent uptake by 2030. This is the environment the cable is being built into. Data consumption is not plateauing.

The risks that could still bite the timeline

None of this is friction-free. A 280MW load is a serious new demand on the grid at a time when renewable generation is being actively debated, and the data centre has drawn opposition on environmental and resource grounds. RNZ’s coverage flags water consent concerns and worries about impacts on sea life and cultural values. Datagrid also released no details on the cable’s cost, its financing, or a construction start date with the July announcement.

An ODT editorial has framed the wider prize as more than NZ$4 billion injected into the regional economy during construction, followed by ongoing data exports. Whether those numbers land depends on clearing the consent and grid hurdles. But the structural logic is sound. For the first time, a third of New Zealand’s landmass gets its own door to the world rather than borrowing the North Island’s. For any data-heavy business weighing where to base itself, that changes the map.

Sources

Community

Join the discussion

Add useful context, ask a good question, or challenge an idea — keep it specific and respectful.

Create a commenter account

Enter the name you want shown publicly and your email. We will email you a password-set link; you cannot comment until you use it.

Your email is used for sign-in and account security. It is not published with comments.

Subscribe for weekly news

Subscribe For Weekly News

* indicates required