Summary

  • Hawaiki Submarine Cable LP operates one cable with two published design capacities: 43.8 Tbps in the reporting around its July 2018 launch, and 67 Tbps on the operator's own network page as retrieved on 24 October 2026.
  • The physical plant has not changed. Three fibre pairs, roughly 15,000 km of route, five landing points: Sydney, Mangawhai Heads in New Zealand, Tafuna in American Samoa, Kapolei in Hawaii and Pacific City in Oregon.
  • In 2022 BW Digital completed its acquisition of Hawaiki Submarine Cable LP, and in the same year announced a second system, Hawaiki Nui. The party that publishes the capacity figure is now also the party selling the next cable — which is the practical reason to care about the arithmetic.

A submarine cable is one of the few pieces of Internet infrastructure where the installer's marketing claim can survive, unamended, for years after the equipment behind it has been replaced. The glass in the water is a passive, expensive, decades-long commitment. The capacity attached to it is a moving statement about the electronics at each shore end, the vendor generation those electronics came from, and the commercial ambitions of whoever signs the purchase order. Hawaiki is a clean case of the two being confused, because both of its published figures are labelled design capacity and neither document explains the other.

Where the number comes from

On 18 July 2018, ITWire reported that Hawaiki Submarine Cable LP had commenced commercial operations for a 15,000 km fibre-optic cable linking Australia, New Zealand, the Pacific and the United States, describing the operator as an independent telecommunications challenger (ITWire, 18 July 2018). That excerpt is one of the few passages in the material behind this article that quotes a publisher directly rather than summarising it.

The next day, ZDNet put numbers on the thing: a USD 300 million build, a system going live between Australia, New Zealand and the United States, construction that began in April 2016, three fibre pairs — two between Sydney and the United States, one from New Zealand to the United States — a design capacity of 43.8 Tbps, and private ownership co-developed by Sir Eion Edgar and Remi Galasso with additional funding from Malcolm Dick. It also named anchor customers: Vodafone New Zealand, AWS, ASTCA and REANNZ (ZDNet, 19 July 2018).

That is good, specific launch reporting, and 43.8 Tbps is the figure most of the world still associates with the system. It is worth being precise about what such a number means at the moment of launch. It is a statement about a configuration: this many fibre pairs, this repeater spacing, this generation of terminal line cards, this modulation scheme, this spectral allocation. Every one of those variables sits at the ends of the cable or inside the repeaters, not in the fibre itself. The fibre's usable bandwidth is enormous; what the operator can actually sell is what its electronics can squeeze into the spectrum, pair by pair.

Now the second document. The operator's own network page — retrieved on 24 October 2026 for this article, in summary rather than as inspected page text — states a total design capacity of 67 Tbps for a system connecting New Zealand, Australia, American Samoa, Hawaii and the US mainland (operator network page). Same cable. Same operator. Same five countries. A design capacity 23.2 Tbps higher.

The honest position is that these two numbers may not be describing the same quantity at all. One may be the capacity available or equipped at ready-for-service in 2018; the other may be the theoretical ceiling of a fuller build-out. Two independent technical references retained for this article make exactly that point: both flag the 43-versus-67 ambiguity and suggest, without confirming it, an initial-or-available versus design distinction (TeleGeography's Submarine Cable Map; Submarine Networks system page). Neither page's full text was available when this was written, which is itself part of the story — more on that below.

So the two figures should be reported, dated, and left standing next to each other. What can be said with confidence is narrower and more useful: the number the world quotes for Hawaiki was never a fixed property of the cable. It was a statement with a date and a source, and the most recent statement carries a different value from the one made at launch.

Three pairs and a moving ceiling

The physical description is stable across a decade of coverage. Three fibre pairs, roughly 15,000 km, five landing countries, operational since July 2018 (ITWire; TeleGeography). Two of the pairs run Sydney to the United States; one runs New Zealand to the United States (ZDNet). That asymmetry is worth pausing on, because it explains a great deal about how capacity claims are made and sold in the region.

A trans-Pacific cable is not a uniform pipe. Each pair is its own optical system with its own terminals at each end. If two pairs terminate in Sydney and one in Mangawhai Heads, then a customer buying capacity in New Zealand is buying into a different slice of the same cable than a customer in Australia, with a different set of upgrade economics behind it. The route length across the Pacific is brutal on signal-to-noise; the longer the span and the fewer the regeneration points, the more the achievable per-pair capacity depends on terminal sophistication rather than on raw spectrum.

The commercial consequence is straightforward. Since 2018 the coherent optics industry has kept shipping new generations of terminal equipment that push more bits into the same pair and the same repeater chain. When an operator adopts one of those generations, the usable capacity of an already-installed cable rises without a single new metre of submarine plant being laid. That is the entire economic appeal of submarine cables as assets: the expensive, slow, permit-bound part happens once, and the capacity on top of it can be refreshed by swapping cards on shore.

Hawaiki's own record includes precisely that move. In November 2020, Telecompaper reported that Hawaiki Submarine Cable would deploy Ciena's WaveLogic 5 Extreme technology, describing Hawaiki as owning and operating the 15,000 km transpacific system connecting Australia, New Zealand and American Samoa (Telecompaper, 23 November 2020). The report is short and forward-looking — a deployment announcement, not a measurement — and it does not state what capacity resulted. But it establishes the mechanism in this specific case: a terminal-technology refresh, applied to existing fibre, announced as a capacity story.

Which means the question 'how much capacity does Hawaiki have?' has at least four defensible answers depending on tense. What was designed in 2016. What was available at ready-for-service in 2018. What was equipped after a 2020 terminal upgrade. And what the owner currently advertises. Only the first is a property of the cable as built. The rest are procurement decisions.

Who decides the figure

Here the story stops being about optics and starts being about corporate control, which is where capacity numbers acquire their real significance.

At launch, Hawaiki was privately owned and co-developed by Sir Eion Edgar and Remi Galasso, with additional funding from Malcolm Dick, and it launched with anchor customers already named (ZDNet). That is a founder-and-anchor-tenant structure: capacity is sold forward to a handful of large buyers, and the published figure serves both as a technical claim and as a signal to the market about how much room remains for everyone else.

In 2022 that structure changed. BW Digital, described as part of the BW Group, completed the acquisition of Hawaiki Submarine Cable LP, bringing the cable into the group's infrastructure portfolio (BW Group notice; Submarine Networks acquisition page). Details are thin in the material retained here: no price, no exact closing date, no list of regulatory approvals. What is clear is the direction of control. The entity that owns the cable, and therefore the entity that decides what gets equipped and what gets advertised, became part of a larger infrastructure group.

That matters for a reason that has nothing to do with secrecy and everything to do with incentives. A founder-financed cable needs to convince a small number of anchor tenants that its capacity is real and abundant. A portfolio operator needs to convince investors that its assets have growth runway. Both of those needs express themselves in a published capacity figure, and they do not necessarily express themselves in the same figure. When the owner changes, the number can change with it — not because anyone is lying, but because the number is answering a different question.

This is why the 43.8-versus-67 gap is worth reporting carefully rather than resolving. The interesting fact is not which figure is right. It is that the same system, described by the same operator, now carries a figure nearly half again as large as the one attached to its launch, and that no document in the retained record explains the difference. A buyer relying on the number has no way to tell whether the increase reflects equipment actually installed, a fuller design that was always intended, or a change in how the term is being used.

The second cable as a decision

There is one more piece of evidence, and it is the most revealing.

In June 2022 — the same year as the acquisition — BW Digital announced a second submarine cable system, Hawaiki Nui (BusinessWire announcement). Trade coverage describes Hawaiki Nui as intended to add capacity and network diversity, and links the acquisition explicitly to the new project (Submarine Networks). The retained summaries describe Hawaiki Submarine Cable LP as owning and operating the original system while BW Digital is the parent company and developer behind the new one.

Read those two facts together and a strategic choice becomes visible. An owner of an existing cable with three fibre pairs and a decade of upgrade headroom has, in principle, two ways to sell more capacity: keep refreshing the terminals on the plant it already owns, or build a new path and sell that. The second option is far more expensive, far slower, and far more visible. It is also the one the new owner announced within months of closing.

That does not mean the existing cable is exhausted. Nothing in the retained record supports such a claim, and the 67 Tbps figure — whatever it measures — suggests the opposite. What it does mean is that the marginal revenue the owner expects to earn is expected to come disproportionately from a new route rather than from further squeezing of the old one. For anyone modelling Pacific capacity supply, that is the single most informative sentence in the entire public record: the owner's own capital allocation reveals what it thinks of the old cable's remaining runway.

For the landing economies, the implications differ by point. American Samoa and Hawaii sit on the route as intermediate landings, and their connectivity value depends on how capacity is priced and partitioned at those landings — a commercial arrangement that no operator press release will settle. Sydney and Mangawhai Heads are the two terminal clusters where the fibre pairs actually end, and where the upgrade economics are concentrated. A second cable that adds network diversity changes the bargaining position of every one of those points, in ways that depend entirely on where it lands and what it connects.

The June 2022 announcement, as retained here, does not say.

What the record does not show

Intellectual honesty requires being explicit about the evidence behind this article, because the gaps are load-bearing.

Three sources survive as provider-reported verbatim excerpts: ITWire's launch report of 18 July 2018, ZDNet's launch-week report of 19 July 2018, and Telecompaper's 23 November 2020 report on the Ciena deployment. Those are quoted text, and they carry the weight of quoted text.

The operator's network page carrying the 67 Tbps figure survives only as a provider-written summary; the page text itself was not directly inspected, and the summary carries no publication date (operator network page). The BW Group notice of the acquisition survives as a summary, as does the BusinessWire announcement of Hawaiki Nui (BW Group; BusinessWire). The two technical-reference pages on route and landing points survive as summaries too, and both explicitly warn that their own page text was unavailable (TeleGeography; Submarine Networks). Where this article states that the operator's page says 67 Tbps, it means that a retrieved summary of that page says so — not that the sentence was read on the page.

Several questions therefore remain open, and they are not decorative. Whether 67 Tbps denotes designed, manufactured or currently equipped capacity is not established. Whether the 43.8 Tbps figure measured the same thing in 2018 is not established. The corporate chain between BW Digital, BW Group and Hawaiki Submarine Cable LP rests on press notices and summaries, not on a partnership register or a regulatory filing. The price, closing mechanics and approvals of the 2022 acquisition are absent.

And the current status, route, timetable and financing of Hawaiki Nui rest on a single June 2022 announcement summary, with no later status document inspected.

One more record is worth stating plainly, because it is unusual to see an organisation be this candid about itself. BTW's own directory entry for this subject notes that Hawaiki Submarine Cable LP appears with two supporting public references and no confirmed operator, and that the entry exists to link public clues to a responsible organisation and its relationships (Hawaiki Submarine Cable LP). In other words, the reference library alongside this article is explicitly a work in progress on exactly the question — who really operates this thing — that a capacity figure quietly depends on. The 2022 acquisition notice answers it at the level of corporate parentage. It does not answer it at the level of who signs the equipment order.

What to ask next

For a network planner, reseller or investor sizing a position in Pacific capacity, the useful questions are not about terabit totals. They are about dates, ownership and equipment.

Ask which figure is being quoted, on what date, and by whom. A capacity number without a date and an owner is a marketing artefact. Ask whether it describes designed capacity, installed capacity or capacity currently available for sale — three different quantities that all get called design capacity in practice. Ask which terminal generation is equipped at each of the five landings and when the last refresh happened, because that, not the fibre, determines what can be sold next year. And ask what the owner is building: an operator that has just announced a second cable has told you something about its expectation of returns on the first.

For a regulator or a public-interest reader, the question is simpler and older. A cable landing is a piece of national infrastructure with a decades-long life, granted on the understanding that it serves the country it touches. The identity of the party that controls it — and the direction in which that party's investment is pointed — matters more than the headline capacity, because control is what determines whether the asset keeps being upgraded or becomes the older half of a two-cable portfolio.

Hawaiki is not a mystery and it is not a scandal. It is a well-built cable with a well-documented launch, a change of owner, a second system on the drawing board and two published capacity figures that nobody has bothered to reconcile in public. The cable is fixed. The number is not. Anyone relying on the number should ask for its date and its owner before signing anything.