9,620 km · 2 Landing Points · 2 Countries · Ready for Service: 2010
| Length | 9,620 km |
|---|---|
| Status | In Service |
| Ready for Service | 2010 |
| Landing Points | 2 |
| Countries | 2 |
| Location |
|---|
| Chikura, Japan |
| Redondo Beach, CA, United States |
Unity is a 9,620 km trans-Pacific submarine cable that lands at Chikura in Japan and Redondo Beach in California. Ready for service in 2010, Unity was Google's first major investment in a submarine cable — the company's entry into owning rather than leasing international fibre capacity. The consortium included Google, NEC, KDDI, Pacnet, SingTel, and Bharti Airtel, with Google as one of six members rather than a lead sponsor. That structure was novel for 2010; by the early 2020s, hyperscaler-led cables (Marea, Equiano) would become routine. Unity was the pattern before the pattern existed.
The cable is paired with EAC-Pacific, a related Pacific cable managed by Pacnet (later acquired by Telstra), with which Unity shares operational infrastructure. In industry catalogues the two are sometimes referred to together as "Unity/EAC-Pacific." For our measurement purposes they operate as a single cable system for the Japan-US route.
Our monitor measures Unity between Redondo Beach on California's coast and Chikura on Japan's Pacific shore. Over 30 days we collected 23 samples in the forward direction with extraordinary consistency:
| Direction | Samples | Min RTT | Avg | Max | StdDev | Hops |
|---|---|---|---|---|---|---|
| Redondo Beach → Chikura | 19 | 104.7 ms | 106.1 ms | 106.2 ms | 0.37 ms | 8 |
| Chikura → Redondo Beach | 4 | 113.3 ms | 116.3 ms | 118.7 ms | ~2 ms | 14 |
A standard deviation of 0.37 ms on the forward direction is among the tightest we have documented on any cable. Nineteen consecutive samples over 30 days landed within a window about a third of a millisecond wide. The minimum of 104.7 ms against a great-circle physics floor of ~83 ms puts the measurement at 1.26× — comparable to PC-1 (1.39×) on a similar route, and tighter than most other Pacific cables we have covered.
Eight IP hops on a trans-Pacific path is remarkable. A packet enters at Redondo Beach, routes through Google's West Coast infrastructure, traverses Unity across the ocean, and exits at Chikura with only six intermediate routers between the two endpoints. This is the signature of a cable where the lead sponsor (Google) operates both endpoints of the traffic it carries — the path is essentially Google → Unity → Google, without third-party transit.
Before Unity, Google bought capacity on commercial submarine cables like any other large internet company. Renting capacity meant predictable pricing but no control — the cable's owners set bandwidth tiers, congestion handling, and upgrade cycles. As Google's Asia-Pacific traffic grew through the late 2000s, leasing became expensive relative to the option of co-investing in construction.
Unity was the first cable where Google paid capital directly for construction and received dedicated fibre pair ownership as a result. The consortium structure meant Google did not bear the full cost — NEC (supplier and investor), KDDI and SingTel (regional operators), Bharti Airtel (Indian operator), and Pacnet all contributed — but Google's participation signalled a strategic shift. A company that had previously treated the internet's physical layer as rented infrastructure was now treating it as its own.
Over the following decade Google would invest in a long series of increasingly hyperscaler-dominated cables: FASTER (2016), Curie (2019), Dunant (2020), Equiano (2022), Nuvem (2026). By 2026 Google operates a larger submarine cable portfolio than most traditional telecoms. Unity was the first step down that path.
Unity and EAC-Pacific are sometimes described as separate cables. In practice they share substantial infrastructure — landing stations, repair arrangements, operational procedures — and for industry catalogues and our dossier database they are treated as a combined system. EAC-Pacific was originally a Pacnet cable built slightly later than Unity but following a parallel trans-Pacific route; when Pacnet was acquired by Telstra in 2015, the two systems' operations consolidated.
This kind of cable-pair consolidation is common in the submarine industry. Cables that land at the same stations and serve similar customer bases often end up operationally merged over time, even when they were built by different sponsors. The merged entity presents a single commercial interface to customers while maintaining two physically independent fibre routes — useful redundancy at minimal operational cost.
Unity was commissioned with six fibre pairs running at the coherent-modulation state of the art in 2010 — approximately 40 Gbps per wavelength on the initial transponders, for a total commissioning capacity of around 4.8 Tbps. Subsequent electronics refreshes have raised per-wavelength rates to 100 Gbps, then 200 Gbps, and likely 400 Gbps on current transponders, pushing aggregate usable capacity significantly higher without modifying the underlying fibre.
The cable's physical body is 15 years old. It is past the halfway mark of its 25-year design life but still delivering measurement-floor latency, as our 0.37 ms standard deviation confirms. Repeaters show no degradation signature. Power-feed continuity through the cable body is intact. What changes over time are the transponders at each end — the submarine infrastructure keeps working.
Unity is a quiet piece of infrastructure history. It is the cable where the hyperscaler-owned submarine model began. Fifteen years after commissioning, it continues carrying Google's Japan-US traffic at near-theoretical physics-floor latency — arguably the clearest vindication of the bet that building cables is more strategic than renting capacity on them.
By the early 2020s, every major hyperscaler operated multiple submarine cables. Microsoft co-invested in Marea with Meta and Telxius. Amazon joined consortium cables across the Pacific. Meta built JUPITER and BIFROST as principal sponsor. The model that Unity introduced — hyperscaler as cable co-owner rather than capacity tenant — became the dominant pattern. New trans-oceanic cables increasingly have at least one hyperscaler in the founding consortium, and many have hyperscalers as principal investors who fund the majority of construction.
The economics that drove this shift are straightforward. A hyperscaler operating data centres on multiple continents pays meaningful capacity fees to traditional carriers; investing the equivalent capital into cable construction returns dedicated fibre at marginal long-term cost. The decision pivots on traffic volume — once a company moves more than a few terabits per second internationally, owning capacity beats renting it.
Live data on the Unity / EAC-Pacific cable page. For other hyperscaler Pacific cables see JUPITER (2020 Google/Meta/Amazon), PLCN (2022 Google/Meta), and APRICOT (2025 Meta/Google/NTT). For other 15-25 year old survivors see PC-1 and FLAG-NAL.
| Status | ✓ Normal |
|---|---|
| RTT | 113.12 ms / base 117.20 ms |
| Last checked | 2026-04-18 20:31 |
Monitored using RIPE Atlas probes. Open monitoring →
Find the actual cable routing distance between any two cities
Open Calculator →