South Pacific Cable System (SPCS)/Mistral

South Pacific Cable System / Mistral is a 7,300 km submarine cable running along the Pacific coast of the Americas, connecting four countries — Chile, Peru, Ecuador, and Guatemala. Ready for service in 2021, it is jointly operated by América Móvil (through its Claro brand) and Telxius, the infrastructure arm spun out of Telefónica. The catalogue name conflates two project identities: SPCS was the original 2021 consortium cable along South America's Pacific shore; the Mistral branding refers to Telxius's northward extension reaching Guatemala. Operationally, they now form one contiguous system.

What makes this cable interesting in our measurements is not a headline latency number but what sits around it — a 7,300 km path delivered at 1.056× the physics floor, and a routing change in early April 2026 that we caught overnight on the reverse direction.

7,300 km at 75.4 ms minimum

Our monitor samples SPCS/Mistral between Puerto San Jose (Guatemala) and Valparaíso (Chile), the two most-separated landing stations. These two points sit roughly 7,000 km apart along the cable body, which is enough distance to make the propagation physics dominate everything else.

The forward direction — Guatemala to Chile — is remarkably clean. Eighteen measurements spanning 30 days sit inside a 4 ms window, with a standard deviation of under 1 ms. The theoretical minimum for a 7,300 km fibre path is 71.4 ms, so at 75.43 ms we are measuring 1.056× the physics floor. That leaves only 4 ms for landing-station gear, coherent transponder decisions, and anything the IP layer adds. It is the signature of a path that the carrier has groomed — a direct fibre route with minimal detour.

The April 11 routing shift

Before April 11, 2026, the reverse direction from Valparaíso to Puerto San Jose showed something quite different: measurements sat at 130–175 ms with 12 traceroute hops, never dipping below 130 ms across three full weeks of observation. Then, over the course of one day, the path changed. From April 11 onward the same probe in Valparaíso reached the same target in Guatemala in 87–90 ms, and the hop count dropped from 12 to 10.

The same RIPE Atlas probe (ID 11257) continued to run throughout. The source, the destination, and our measurement logic were unchanged. What changed was somewhere between Chile and Guatemala — either a BGP policy adjustment by one of the carriers along the return path, or an operator-side re-grooming that replaced two hops and ~43 ms worth of extra fibre with a more direct return.

Before the shift, the reverse path looked like it was bouncing through a carrier hub — plausibly a return via a US exchange point in Miami or Los Angeles, which is a common pattern for Latin American traffic. After April 11, it looks like the carrier enabled a symmetric return on the same SPCS/Mistral fibre body. The forward direction was never 43 ms off the floor, so both halves now run near physics.

An owner-operated Pacific backbone

SPCS/Mistral is the Pacific-coast counterpart to AMX-1, the 17,800 km cable that covers the Atlantic and Caribbean coasts of the Americas. AMX-1 runs from Jacksonville down through Mexico, Central America, the Caribbean islands, and northeastern Brazil; SPCS/Mistral runs from Guatemala down through Ecuador, Peru, and Chile on the Pacific side. Between the two cables, América Móvil operates a private Latin American backbone that crosses both coasts without depending on competitors' submarine capacity for regional traffic.

The ownership split on SPCS/Mistral is different from AMX-1, though. AMX-1 is wholly América Móvil; this cable is shared with Telxius. Telxius was carved out of Telefónica in 2016 and has built out one of the larger pure-infrastructure cable portfolios in the Atlantic — Marea (with Meta and Microsoft), Brusa, Dunant (with Google), and this Pacific system. For Telxius the Pacific presence closes a gap: their Atlantic portfolio is strong, but they needed a Pacific route to offer global wholesale capacity without leasing from carriers in Asia.

Five landing stations

The Arica landing deserves a note. Arica sits a few kilometres from the Peruvian border at Chile's northern tip — historically a redundancy point, because a cable body on the Pacific floor can be serviced from either the Peruvian or Chilean side depending on where a fault sits. Two Chilean landings also let the cable deliver traffic into both the Santiago metropolitan area (via Valparaíso) and the mining-industrial north (via Arica) without needing to backhaul across Chilean terrestrial networks.

Six fibre pairs, 132 Tbps design capacity

SPCS/Mistral was built with six fibre pairs and a design capacity of 132 Tbps according to Submarine Networks' published specs. That is a moderate number by 2021-era standards: Marea had eight pairs (2018), Apricot has twelve pairs (2025), and Medusa has twenty-four pairs (2026). SPCS/Mistral's six pairs reflect the reality that Latin American west-coast traffic demand is lower than trans-Pacific or trans-Atlantic flows — the cable was sized for actual regional need rather than speculative capacity.

Like all modern submarine systems, capacity on paper is not fixed. The wet plant — fibre and repeater hardware on the sea floor — is good for 25 years. The dry plant — coherent transponders at each landing station — is upgraded every five to seven years, and each upgrade roughly doubles per-wavelength throughput. The six fibre pairs laid in 2021 will almost certainly be lit at 300–500 Tbps before the cable retires.

Why a near-floor measurement matters

A 1.056× physics-floor measurement on a 7,300 km path is a strong positive signal. It says: the carrier is running this cable the way it was designed to be run. Traffic between Central America and Chile is not detouring through a North American exchange point, not being hairpinned through a congested peering link, not bouncing off a saturated wavelength. The cable is doing the job its geography makes possible.

Compare the architectural equivalents we have measured on other routes:

• AMX-1 Jacksonville → Rio: 118.7 ms over a great-circle 7,300 km. Coastal cable, adds ~4,800 km of fibre for multi-country coverage — 1.66× great-circle.
• SPCS/Mistral Guatemala → Valparaíso: 75.4 ms over a cable-path ~7,000 km. Coastal, but the Pacific coast of the Americas is close to a straight line, so great-circle and cable-path largely agree — 1.056× floor.
• PC-1 Seattle → Tokyo: 103 ms over 21,000 km. Trans-Pacific hyperscaler-era, 1999.

SPCS/Mistral's measurement sits at the lowest-overhead end of that spectrum because the Pacific American coastline is geographically cooperative: no big detours required to land at every intermediate country.

What our data proves

• Guatemala → Chile at 75.4 ms with 0.95 ms variance across 18 measurements. Forward direction runs at 1.056× physics floor on a 7,300 km cable path.
• Reverse direction changed on April 11, 2026. The same probe dropped from 131 ms / 12 hops to 88 ms / 10 hops overnight, indicating a carrier-side route change that brought return symmetry to the cable.
• Ownership split. Joint América Móvil (Claro) and Telxius build — Latin America's largest operator and Telefónica's infrastructure spin-out sharing a single Pacific asset.

Try it yourself

Live measurements on the SPCS/Mistral cable page. Compare with AMX-1 (the Atlantic-coast sibling) and South American Crossing (the older loop around the entire continent). The three together sketch how Latin American submarine infrastructure has evolved — from one-cable-around-everything in 2000 to dual-coast dedicated backbones by the mid-2020s.