Bulikula

Bulikula is a 21,600 km submarine cable system lit for service in 2026, connecting a geographically scattered set of Pacific landings: Kapolei in Hawaii, Piti and Tinian in the Mariana Islands, Natadola and Suva in Fiji, and Faratea, Papenoo, and Mitirapa in French Polynesia. The cable's shape looks less like a trans-Pacific trunk and more like a hub-and-spoke tree that stitches the northern Pacific islands into a single fibre body. With 16 fibre pairs across its branches, Bulikula carries capacity between island communities that previously depended on older, thinner cables or on transit through continental landings in North America or Australia.

Bulikula is one of a growing cohort of Pacific-island-focused cables being commissioned in the mid-2020s. Alongside Tonga Cable, Gondwana-2, Southern Cross NEXT, and Hawaiki, it is part of a deliberate buildout to provide Pacific island nations with direct high-capacity fibre rather than the satellite-plus-legacy-cable arrangements many of them depended on through the 2010s.

217 ms across the Pacific

Our monitor measures Bulikula between Piti in Guam and Faratea in French Polynesia — two endpoints on opposite ends of the cable's reach. Over 30 days we collected 32 samples in both directions:

The forward direction (Piti to Tahiti) is remarkably stable on most days, with clean 217 ms minimums that cluster tightly across clean 24-hour windows. The occasional excursions to 320–440 ms on isolated days reflect brief congestion or path instabilities at the intermediate island landings, but they resolve back to 217 ms baseline quickly.

Light in submarine fibre has a theoretical round-trip minimum of 211 ms for a 21,600 km path. We measure 217.1 ms — that is 1.03× the physics floor. Remarkable for a cable that winds through six+ landings across three island groups. The packet likely travels the full cable length (Guam down through Hawaii/Kapolei, across to Fiji, onwards to French Polynesia) along Bulikula's branches, hitting each intermediate landing as a brief routing handoff.

Three scattered archipelagos, one cable body

The landing structure is what makes Bulikula architecturally unusual. Most cables we have covered have landings either along a continental coast (AMX-1, WACS, Medusa) or span two continents with one or two intermediate island landings (Marea, Equiano, Nuvem). Bulikula has no continental landings outside of Hawaii — it is purely a Pacific-island infrastructure project.

This reflects the cable's purpose. Pacific Island Countries and Territories (PICTs) have dramatically different connectivity needs than continental internet users. Populations are small (Fiji's 900,000 is one of the larger Pacific nations), but geography makes inter-island communication disproportionately difficult. Before the current cable buildout, a video call between Fiji and French Polynesia would typically route through Los Angeles or Sydney — a trip of 30,000+ km for what is physically an 8,000 km direct path. Bulikula collapses that kind of routing.

The Tinian landing is worth noticing. Tinian is a small island in the Northern Marianas with a population under 3,000. The cable landing represents significant infrastructure investment for an island of that scale — the economics only make sense as part of a larger regional network where Tinian serves as a strategic Pacific midpoint for routing between the Northern Pacific (Hawaii, Guam) and the Southern Pacific (Fiji, French Polynesia).

16 fibre pairs for small markets

Bulikula has 16 fibre pairs across its trunk. The Submarine Networks catalogue lists the cable's design capacity at 16 Tbps — 1 Tbps per fibre pair, which is below the modern ceiling. For comparison, APRICOT's 12 pairs deliver 290 Tbps (24 Tbps per pair) and Medusa's 24 pairs deliver 480 Tbps (20 Tbps per pair).

Why does Bulikula sit below the ceiling? Several possibilities. First, Pacific-island markets have much smaller total demand than the Asia-Europe or Asia-Americas trunks, so overprovisioning capacity has less commercial return. Second, the listed 1 Tbps per pair may reflect the initial launch configuration — future electronics upgrades can raise per-pair capacity substantially without physical changes to the fibre, which is the standard submarine cable lifecycle. Third, branching architectures with many landings incur signal-integrity trade-offs compared to straight point-to-point trunks, which can limit practical per-pair capacity.

Whatever the reason, Bulikula's 16 Tbps initial capacity is still a transformational increase for the Pacific island communities it serves. Many of these territories previously depended on cables with under 100 Gbps of available capacity — less than a rounding error in Bulikula's design.

Eight days per cable fault

Pacific island cables have a different repair profile than continental systems. When a cable breaks on the US-Europe route, a cable repair ship can typically be on-station within 48 hours. Pacific island cables, by contrast, serve regions where cable repair ships are often days away — both the Hawaiian and Asian repair fleets have finite capacity, and weather in the Pacific can delay operations by a week or more.

Bulikula's 16 fibre pairs and multi-landing architecture provide resilience against this. If a single undersea segment fails, traffic between Fiji and French Polynesia can be rerouted through Hawaii or Guam via the remaining paths. A single cable fault no longer means weeks of disconnection for affected islands — traffic simply takes a longer route through other branches of the same cable.

This design philosophy — capacity traded for redundancy — is central to the current Pacific cable buildout. Each new cable serving the region is designed to carry traffic independently of the others, so that the total regional connectivity does not depend on any single system remaining healthy.

What our data proves

• Bulikula delivers 217.1 ms between Guam and French Polynesia, 1.03× the physics floor. Near-theoretical performance across 21,600 km of fibre winding through multiple island landings.
• Both directions agree. Forward and reverse minimums land within 3 ms of each other — consistent with symmetric routing across the cable's trunk between these two endpoints.
• Occasional spikes resolve quickly. Maximum RTT excursions to 400+ ms on isolated samples reflect brief congestion at intermediate landings, but the baseline returns to 217 ms within the same 24-hour window.

Bulikula is part of a generation of Pacific-focused submarine cables that is quietly transforming the connectivity options available to small island economies. The cable's 2026 commissioning is one data point in a multi-year buildout; its effect on Pacific connectivity will accumulate over the following decade as more cables on similar routes light in parallel.

Try it yourself

Live data on the Bulikula cable page. For context on Pacific cables, see Tonga Cable (single-cable island), JUPITER (trans-Pacific hyperscaler), and our research article on Cook Islands connectivity via Manatua.