Monet

Based on 38 RIPE Atlas measurements from GeoCables monitoring infrastructure, March–April 2026.

Monet is a relatively short submarine cable — 10,556 km of fibre connecting Boca Raton, Florida to two landings in Brazil (Fortaleza and Santos) — that first entered service in 2017. It is also one of the cleanest performers in our entire monitoring queue. Across 38 measurements from Boca Raton to Santos, the minimum round-trip we observed is 117.01 ms, the average is 117.66 ms, and the standard deviation across all measurements is a remarkable 2.38 ms. The theoretical physics floor for that 10,556 km path is 103.31 ms. Monet measures at 1.133× floor, which is tight — but the real story is the standard deviation.

Routing stability at this level — a cable whose measured RTT varies by less than 2.5 ms across dozens of samples — is unusual. It suggests a path that transit providers have converged on completely, with no competitive alternative path offering similar latency.

What Monet is

Monet's ownership is the kind of mix that makes 2017-era cables architecturally interesting. The consortium has four members: Algar Telecom (a mid-sized Brazilian integrated operator, headquartered in Minas Gerais), Angola Cables (the Angolan state-owned telecom with a strong Atlantic-crossing strategy), Antel Uruguay (the Uruguayan state carrier), and Google. It is the earliest Google-funded submarine cable we've covered on the Latin American side, predating both Nuvem (Google's 2025 Pacific-Atlantic cable) and Meta-era Latin American systems like Malbec (2021).

Three landings, two of them in Brazil, one in the US — this is a purpose-built US-Brazil cable. Unlike GlobeNet (which we published earlier in this session: 7 landings across 5 countries, with Caribbean branches and a Bermuda loop), Monet does not try to serve multiple regional markets. It does one thing — move traffic between Florida and the two largest Brazilian internet hubs — and it does it with the tight geometry of a purpose-built point-to-point system.

The 2.38 ms standard deviation

On a baseline RTT of 117 ms, a standard deviation of 2.38 ms means that 95% of measurements fall within roughly 4.8 ms of the average. The maximum we observed across all 38 samples is 129.56 ms — only 12.5 ms above the floor. There are no re-routing excursions in this data, no BGP flap artefacts, no congestion-driven detours. The transit providers carrying Boca Raton-to-Santos traffic have converged on a single stable path through Monet, and that path just stays put.

For reference, many of the cables we've published show standard deviations of 20–100 ms or higher on similar-length paths. SACS — another South Atlantic cable — runs near-floor but with a spread three times wider. TAM-1, the newer Trans Americas pure-play, shows 8.9 ms variance that we attributed to first-year adoption curve. Monet, now eight years old, has passed through that curve and settled into the pattern of a mature, well-routed cable.

Google's role

Google's participation in Monet predates the peak of its direct-investment era. By 2017, Google had already funded Unity (2010) and Unity/EAC-Pacific (we published earlier), but Monet was one of its earliest explicit multi-party submarine-cable consortia — where Google took a capacity stake alongside traditional telecom partners rather than funding a whole cable alone. The pattern that became standard by 2022 — Google building cables like Equiano as sole funder — was less established in 2017.

From a performance standpoint, Google's presence likely contributes to the tight routing pattern. Google's internal traffic between Google Cloud's São Paulo region and its US data centre network follows predictable paths through the cable, and Google's peering relationships with Brazilian transit providers are well-optimised. That's exactly the kind of stable, large-volume flow that would drive the converged routing pattern we measure.

Algar and Angola Cables contribute regional commercial depth; Angola Cables in particular operates the SACS cable to Africa, which means Monet's Brazilian endpoints plug into a trans-Atlantic African network through the same operator. This is how modern cable consortia work: each partner brings a different adjacent market, and the consortium is the meeting point.

The Latin American traffic corridor, 2026 edition

Eight years after launch, Monet sits in a US-Brazil corridor that has become one of the most heavily served submarine routes in the world. AMX-1 (2014), GlobeNet (2000), South American Crossing (2000), and now TAM-1 (2026) all compete for the same broad class of traffic. What Monet offers that the older cables don't is the capacity density of a modern optical system, a simpler architecture without legacy Caribbean-loop overhead, and — as our measurements show — a remarkably stable performance envelope.

The commercial implication is that Monet's co-owners have something specific to sell: Florida-to-São Paulo transit at 117 ms floor with tiny variance, on a Google-backed infrastructure, marketed to customers (enterprise, cloud-provider transit, CDN origins) who care about both latency and predictability. On the slightly older corridor cables, the same route might clear at similar latency on a good day, but with much wider variance when routing conditions shift.

Why Brazil is the US internet's largest Latin American counterparty

The volume of traffic that Monet and its cousin cables carry reflects a specific commercial reality. Brazil is, by a considerable margin, the largest internet market in Latin America — both by user count (over 180 million active internet users) and by economic activity of the digital economy. São Paulo specifically hosts the Latin American data-centre presence for every major hyperscaler: AWS, Google Cloud, Microsoft Azure, and Oracle Cloud all operate significant regions there, and the dependency between those regions and their US primary data centres is enormous. Every piece of content, every API call, every database replication between Brazil and the global cloud footprint crosses a cable like Monet.

This makes the corridor unusually valuable relative to other Latin American routes. Argentina, Colombia, Chile, and Mexico all have meaningful data-centre and internet footprints, but none approaches Brazil's scale. For a Google-invested cable in 2017, picking the Boca Raton–São Paulo pair was less a contested architectural choice than an obvious one: go where the traffic is. And the traffic was, and remains, concentrated on that specific intercontinental pair.

What we're watching

Two things. First, whether the floor stays at 117 ms. An eight-year-old cable is still young in submarine terms, but drift in the floor metric would be the earliest signal of capacity-level saturation or optical-path degradation. Second, whether a reverse direction appears in our probe rotation. Right now we measure only Boca Raton → Santos. A São Paulo or Rio-based probe would let us check symmetry and expose any asymmetric routing preferences between US and Brazilian transit providers on the same cable. Given how tight the forward direction is, asymmetric behaviour in the reverse would be meaningful — it would mean different operators are making different routing choices despite sharing the same underlying Monet capacity.