Pacific Light Cable Network (PLCN)

The Pacific Light Cable Network (PLCN) is an 11,806 km trans-Pacific submarine cable lit for service in 2022. It connects El Segundo on the southern California coast with Toucheng in Taiwan and Baler in the Philippines, operated by a consortium that includes Google and Meta alongside regional Asian partners. With six fibre pairs and a design capacity of 24 Tbps, PLCN was built as hyperscaler infrastructure: a private pipe for Google and Meta data between their West Coast US facilities and their Asian data centres.

PLCN occupies an unusual place in recent submarine cable history. It was originally proposed in 2016 as a four-landing cable: Los Angeles, Taiwan, the Philippines, and Hong Kong. The Hong Kong segment was laid. Cable, landing station, termination — all built. But in 2020, after a multi-year national-security review by US regulators (Team Telecom, which advises the FCC on foreign cable applications), Google and Meta requested and received approval to activate PLCN only to Taiwan and the Philippines, leaving the Hong Kong segment dark. It is one of the most high-profile examples of a trans-Pacific cable's architecture being reshaped mid-project by regulatory decisions rather than by engineering.

118.6 ms: nearly physics floor

Our monitor measures PLCN between El Segundo and Baler. Over 30 days we collected 29 forward-direction samples and 7 reverse. The forward direction is remarkably stable:

Light in submarine fibre travels at about 204,500 km/s. A round-trip through 11,806 km of glass has a theoretical minimum of 115.4 ms. We measure 118.6 ms — that is 1.03× the physics floor. PLCN's forward direction delivers close to the absolute limit of what physics allows for its length. The 8-hop count is also notable: a very clean path with minimal intermediate routing between the Los Angeles region and a Philippine endpoint.

Outliers in the forward data (the 291 ms spike on April 2, and the jitter around 119 ms baseline) are few. On a typical day, every sample lands within 0.2 ms of the minimum. That is the signature of a dedicated hyperscaler pipe — owners routing their own traffic directly across their own fibre pairs.

64 ms of reverse asymmetry

The reverse direction — Baler to El Segundo — measures 183.0 ms at its fastest and takes 16–17 hops (double the forward). Reverse traffic is plainly not using the same route. 183 ms round-trip corresponds to about 18,700 km of fibre, substantially longer than PLCN's 11,806 km, suggesting the return path uses a different cable entirely or transits multiple systems before reaching the US.

This is a smaller asymmetry than we documented on JUPITER (200 ms gap) or EIG (130 ms), but still significant. The pattern is consistent with what we see across other hyperscaler Pacific cables: owners route forward traffic tightly through their own cable, while return traffic from the far end passes through a carrier-chosen path that may prefer a shorter or cheaper alternative like JUPITER, FASTER, or one of the Asia-to-Americas routes through Japan.

Why Hong Kong didn't happen

PLCN was designed in 2016, when a Hong Kong landing was uncontroversial. The city was one of the two largest internet exchange hubs in Asia (the other being Singapore), and nearly every trans-Pacific cable either landed there or had plans to. Google, Meta, and their Chinese partner Pacific Light Data Communication (PLDC) proposed PLCN as a standard four-landing cable.

Between 2018 and 2020, the US regulatory environment for trans-Pacific cables shifted. Team Telecom — an interagency body that reviews foreign-owned or foreign-landing cables for national-security implications — began recommending against Chinese-controlled landings in the US, and by extension against US-landed cables that also touched jurisdictions with significant Chinese network visibility. A series of cable applications were delayed, withdrawn, or restructured during this period.

For PLCN, the resolution was to activate only the portions without Hong Kong involvement. The cable body extending to Hong Kong remained on the seafloor — submarine cables cannot be practically retrieved once laid — but the Hong Kong landing station was never turned up as part of the operational system. Google and Meta proceeded with El Segundo, Taiwan, and the Philippines, with PLDC eventually exiting the project.

The technical consequence for our data is that PLCN's Taiwan and Philippines landings carry the traffic the cable was designed for. Our El Segundo → Baler measurement shows the cable performing exactly as intended on the routes that were allowed to light.

Three-country cable, eleven-thousand km trunk

Three-landing cables are uncommon in the modern generation. Most new Pacific cables (JUPITER, BIFROST, APRICOT) have five to eight landings to serve multiple regional markets. PLCN's narrower design reflects its role as a private hyperscaler pipe: Google and Meta wanted Asian capacity, and they picked Taiwan and the Philippines as endpoints that could house their own data-centre backhaul and peering infrastructure without the complications of additional landings.

The Philippine landing at Baler, on Luzon's Pacific coast, is particularly strategic. The area has been developed as a hyperscaler-friendly cable hub, with Meta, Google, and their Philippine partners operating cable landing stations and backhaul connectivity into the Manila data-centre cluster. PLCN's Baler landing is part of Philippine government plans to develop the east coast of Luzon as an alternative to the crowded western cable landings around Batangas.

Six fibre pairs, 24 Tbps

PLCN was designed with six fibre pairs and a total capacity of 24 Tbps — 4 Tbps per pair at the original specification. This is modest by 2025 standards (APRICOT has 12 pairs and 290 Tbps; BIFROST has similar density), but it reflects both the 2016-era design and the single-sponsor nature of the capacity. Google and Meta each own one or more pairs and do not need to sell wholesale capacity to carriers — so there is less pressure to maximise per-pair density.

Like all submarine cables, PLCN can be upgraded on the electronics side as coherent transponder technology improves. With 400 Gbps per wavelength and extended C-band in the late 2020s, per-pair capacity could approach 10–15 Tbps without touching the fibre. But the cable's inherent value to its owners is not raw capacity — it is owned capacity, dedicated to their own traffic, with predictable latency and no carrier middlemen.

What our data proves

• PLCN delivers El Segundo → Baler at 118.6 ms, 1.03× the physics floor. The cable performs essentially at the theoretical limit of glass-fibre latency for its 11,806 km length.
• Forward performance is remarkably tight. 29 samples over 30 days in a ~1 ms band around the minimum. This is a well-groomed hyperscaler route.
• The return direction uses a longer, multi-hop path. 183 ms reverse latency suggests Philippine-to-US return traffic is routed via another system (likely through Japan or Singapore), not back along PLCN.

PLCN is a reminder that submarine cable architecture is shaped by more than physics, economics, and engineering. Regulation matters. A cable can be fully built and still not operate the way its designers planned. Our measurements show the cable working exactly as it was approved to work — and the performance is excellent for the segments that were lit.

Try it yourself

Live data on the PLCN cable page. Compare with other Pacific cables: JUPITER (dramatic asymmetry), APRICOT (intra-Asian hyperscaler, symmetric), and BIFROST (southern-corridor Asia-to-Americas).