America Movil Submarine Cable System-1 (AMX-1)

The America Movil Submarine Cable System-1 (AMX-1) is a 17,800 km submarine cable running down the Atlantic and Caribbean coasts of the Americas, connecting eight countries — the United States, Mexico, Guatemala, Colombia, the Dominican Republic, Puerto Rico, Costa Rica, and Brazil. Ready for service in 2014, it is owned by América Móvil (Latin America's largest telecom group) as a unified backbone across the regions where its subsidiaries operate. Unlike the hyperscaler cables that sprint across an ocean as direct point-to-point links, AMX-1 is a coastal cable: it hugs the Caribbean and Atlantic shores, landing at a dozen locations on its way from Jacksonville, Florida down to Fortaleza in northeastern Brazil.

That architectural choice — a long, many-station cable following the coastline rather than a short, direct cable cutting across open ocean — produces a latency signature you can read straight off our measurements.

17,800 km on paper, roughly 12,000 km of fibre in use

Our monitor measures AMX-1 between Jacksonville and Rio de Janeiro — the two longest-separated landing stations of the eight. Over 30 days we collected 16 forward-direction samples, with remarkable consistency:

The forward direction is astonishingly tight — sixteen measurements spread over a 2 ms window. The hop count is also stable at 14. This is a cable-path that carriers are actively using and have groomed for consistent performance.

Here is where AMX-1's architecture becomes visible. The great-circle distance from Jacksonville to Rio de Janeiro is approximately 7,300 km. The cable itself is 17,800 km total length, but that includes every landing between the two — Mexico, Central America, Colombia, the Caribbean islands, Brazil's northern coast. The path actually used between our two monitored endpoints is neither the direct 7,300 km great-circle nor the full 17,800 km cable — it is somewhere in between.

118.7 ms round-trip corresponds to about 12,100 km of fibre traversed. That is 1.66× the great-circle distance — AMX-1's coastal routing adds roughly 4,800 km of fibre compared to a straight Atlantic crossing. In exchange, the cable serves eight countries with direct landings instead of just two endpoints.

The coastal-cable tradeoff

Submarine cable designers face a choice: do you optimise for the shortest path between two endpoints, or do you land in many places to serve multiple markets? The two goals are in tension. Landing the cable means bringing it up into shallow water, through a shore approach, and into a cable station — each landing adds 50–200 km of fibre depending on the coastal geography, and it introduces a brief routing handoff where traffic can be groomed or dropped.

A direct trans-Atlantic cable like Marea (Virginia Beach ↔ Bilbao, 6,600 km) takes the shortest practical route between the two endpoints. A coastal cable like AMX-1 prioritises coverage: one contiguous system serves every subsidiary of América Móvil's business, from Florida down to Brazil, without the company needing to lease capacity from competitors for intra-region traffic.

The Jacksonville → Rio de Janeiro measurement shows what that tradeoff looks like in milliseconds. A direct Atlantic cable between those two cities could deliver 72 ms RTT (at 1.0× physics floor for 7,300 km). AMX-1 delivers 118.7 ms — about 46 ms slower. In exchange, AMX-1 also delivers direct connectivity to Mexico City, Guatemala City, Bogotá, Santo Domingo, San Juan, and San José on the same fibre pair — none of which a direct Virginia-Beach-to-Rio cable could serve.

Eight countries on one cable

América Móvil (which operates under brands including Claro, Telcel, and Telmex) is the largest fixed-line and mobile operator across Latin America. Before AMX-1, its regional traffic had to transit through either US-based international hubs or via competitors' cables to move between its own national subsidiaries. Owning a private backbone that connects all of its operating regions on a single cable body collapses that dependency.

The cable's geopolitical positioning is also worth noticing. By landing in the Dominican Republic, Puerto Rico, and Colombia, AMX-1 provides Caribbean nations and Colombia with a direct path to both Florida and Brazil without going through US border infrastructure. For Colombian and Caribbean telecoms, that changes the commercial dynamics of trans-American internet connectivity — reducing the transit fees paid to US wholesale providers.

The 37 ms asymmetry

Forward 118.7 ms vs reverse 152.8 ms is a 34 ms asymmetry — smaller than we have seen on JUPITER (200 ms) or EIG (130 ms) but still meaningful. The hop count differs too: 14 forward, 18 return. Both observations point to the return path using a somewhat different route than the forward — likely one or more extra transit hops through US peering infrastructure that are not on the direct Brazilian-outbound path.

34 ms of extra return-path latency corresponds to about 3,500 km of additional fibre. That might be a routing loop through Miami or Houston, or it might be a return path that goes via a different Caribbean landing. Without visibility into the carrier's internal path selection, we cannot say for sure. But the pattern is stable — every return sample we captured over 30 days shows this 34 ms gap — so it is the steady-state routing policy, not a transient.

AMX-1 in 2026

The cable has been operational for over a decade. Like most cables of its generation, it has been upgraded on the electronics side several times — the submarine body unchanged, but the transponders at each landing station stepped through higher-capacity coherent modulations. That is the submarine infrastructure lifecycle: 25-year design life on the wet plant, 5–7 year refresh cycle on the dry electronics that determine actual capacity.

A decade ago, AMX-1 was a novel piece of infrastructure: a single telecom group building its own cross-regional backbone. Today that model is routine — every major Latin American carrier operates at least one such cable, and hyperscalers have added their own regional systems (Firmina, Curie) with similar coastal designs. AMX-1 was an early example of a cable built for operational independence rather than for wholesale capacity — a pattern that has since become the norm.

What our data proves

• AMX-1 delivers Jacksonville → Rio de Janeiro at 118.7 ms with 2 ms of variance across 16 measurements. A decade after commissioning, the cable's forward-direction performance is clean and consistent.
• The coastal-cable design adds about 4,800 km of fibre to the great-circle distance. That is the cost of serving eight countries with direct landings instead of two endpoints with a direct cable.
• Reverse direction takes a 3,500 km longer path. The asymmetry is 34 ms — smaller than hyperscaler trans-ocean cables but still a visible fingerprint of BGP policy shaping which direction uses which route.

Try it yourself

Live data on the AMX-1 cable page. Compare with direct trans-oceanic cables like Marea (1.95× floor on a straight Atlantic) and Equiano (2.5× floor on Portugal → South Africa). AMX-1's coastal architecture is a useful foil for understanding why the physics floor alone does not tell you how a cable will perform — the shape of the route matters just as much as the length.