South Atlantic Cable System (SACS)

The South Atlantic Cable System (SACS) is the first submarine cable to connect Africa and South America directly. Before SACS came online in 2018, internet traffic between Angola and Brazil — two countries separated by roughly 6,000 km of ocean — had no practical option except to route through Europe. A packet from Luanda to Rio de Janeiro would hop north through European submarine cables, cross the Atlantic on a North Atlantic system, and descend to Brazil. Total round-trip: typically 300+ ms. SACS collapsed that detour. It is 6,165 km long, lands in Sangano (near Luanda, Angola) on one end and Fortaleza, Brazil on the other, with an intermediate branching point serving Fernando de Noronha, a small Brazilian archipelago in the mid-Atlantic.

SACS was built and is operated by Angola Cables, a carrier majority-owned by Angola Telecom. Unlike hyperscaler cables sponsored by Google or Meta, SACS was built as a national telecom investment — a strategic statement that Africa's connectivity no longer needed to run through former colonial hubs in Europe. The cable has four fibre pairs, commissioned with a 40 Tbps design capacity that upgrades have since pushed higher.

62.4 ms: sitting on the physics floor

Our monitor measures SACS between its two end landings — Fortaleza in Brazil and Sangano in Angola. Over 30 days we collected 25 forward-direction samples and 3 reverse. The forward direction is extraordinarily consistent:

Light in submarine fibre moves at about 204,500 km/s. A 6,165 km round-trip fibre path has a theoretical minimum RTT of 60.3 ms. We measure 62.4 ms — just 1.03× the physics floor. That is as close to theoretical limit as we have documented on any cable.

The forward direction takes only five IP hops. Five is remarkably few for a trans-ocean path — it implies a very clean route with minimal intermediate routers. The probe near Fortaleza reaches an Angola Cables-operated endpoint (197.149.149.17) within that minimal chain, suggesting Angola Cables operates both endpoints directly without long terrestrial backhaul at either end.

For comparison, Equiano (Portugal ↔ South Africa, similar length at 6,100 km) measures at 2.5× the physics floor and 13 hops. Marea (Virginia Beach ↔ Bilbao, 6,600 km) measures 1.95× at 13 hops. SACS is almost twice as fast as either, measured from our probes. That is not because the glass is faster — it is because the monitored path uses a minimal-backhaul, direct cable-station-to-cable-station route, with both endpoints operated by the same carrier.

The 46 ms return-path detour

The reverse direction — Sangano toward Fortaleza — measures 108.2 ms, 46 ms slower. Eleven hops rather than five. Different probe, different target, different route. This is not the same path.

108 ms round-trip corresponds to about 11,050 km of fibre. That is nearly double the length of SACS itself, which means the return path is probably not using SACS at all. Likely route: Sangano → up the West African coast via WACS to Europe → across the Atlantic on a Europe-Brazil cable (ellaLink or similar) → south to Fortaleza. Two cables, three continental handoffs, 11 hops total.

Why would Angola Cables' own probe choose the long way around to reach Brazil? Three possible reasons:

• The probe is not on Angola Cables' production network. The 108 ms measurement uses 60219, a probe that likely lives on a different Angolan carrier. That carrier may not have SACS peering set up on its outbound route.
• The target is reached via a Brazilian carrier whose best route is European. The target IP (200.160.2.3) belongs to NIC.br, the Brazilian internet registry. NIC.br's upstream peering is heavily European — European carriers reach Brazil via direct Atlantic cables, so NIC.br's announced path may preference those cables over SACS.
• Historical routing inertia. Many traffic engineers on both continents grew up routing Africa-Brazil through Europe. Even years after SACS lit, production routing tables carry legacy preferences that take time to update.

The asymmetry is steady across 30 days of measurement. This is not a transient routing glitch — it is the steady-state policy of the operators involved. SACS exists and delivers near-physics-floor performance from Fortaleza; the return trip is still routed the old way.

Why a direct Africa-Brazil cable mattered

SACS was the culmination of a decade of capacity planning by Angola Cables. When construction began in 2016, the commercial case was not obvious: African carriers historically had plenty of capacity to Europe and only secondary demand for Brazilian peering. But three factors converged.

First, data-centre footprint in Brazil grew rapidly in the mid-2010s, with hyperscalers and local operators opening large facilities in Fortaleza and São Paulo. African enterprises serving Brazilian or Portuguese-speaking customers needed low-latency paths that Europe-transit could not provide.

Second, Angola Cables wanted to become a regional transit hub, not just a national operator. Owning a trans-Atlantic cable gave it wholesale capacity to sell to other African carriers — EASSy, WACS, and inland neighbours could buy Atlantic capacity from Angola rather than lease it from European operators.

Third, strategic redundancy. African internet traffic that passes through Europe inherits European vulnerabilities — cable faults in the Mediterranean, regulatory or geopolitical friction, and shared infrastructure bottlenecks. A direct Africa-South America route gave African operators a physically independent Atlantic option.

SACS pairs naturally with Equiano (Google, 2022, Portugal ↔ South Africa) and WACS (2012, Lagos ↔ Seixal). Together they form a West African submarine mesh that no longer depends on any single European landing. That is the structural shift Angola Cables underwrote by building SACS.

Fernando de Noronha

SACS has an intermediate branching point serving Fernando de Noronha, a small Brazilian archipelago 350 km off the Brazilian coast. This was not a commercial decision. Fernando de Noronha is a UNESCO Heritage site and a Brazilian national park with a permanent population of roughly 3,000. Its inclusion as a branching point is partly practical (the cable needed a mid-Atlantic repair handoff point) and partly symbolic (Brazil gets a direct fibre touch-down on one of its most iconic Atlantic territories).

The spur is small — a tiny fraction of the cable's total capacity — but it gives the islands direct terabit-scale connectivity, which before 2018 they did not have. For the 3,000 people living on a UNESCO site in the middle of the South Atlantic, being on the first Africa-Brazil cable means their internet now has the same latency floor as any major port on the continent.

What our data proves

• SACS operates at 1.03× the physics floor on the Fortaleza → Sangano direction. 62.4 ms across 6,165 km of trans-Atlantic fibre — among the tightest measurements in our entire dataset.
• The reverse direction uses a Europe-transit path. 108 ms and 11 hops correspond to the old Africa → Europe → Brazil routing that SACS was built to replace — still dominant in production traffic eight years after the cable's commissioning.
• Minimal-hop architecture. 5 IP hops on the forward path is remarkable for a trans-oceanic cable. It reflects single-carrier operation at both endpoints and minimal intermediate routing.

SACS is one of the quietest, most effective submarine cables in our monitoring set. It does the thing its builders wanted — deliver Africa-Brazil traffic at physics floor — and it has done it reliably for eight years. The only remaining question is how many years of production routing policy it will take for the rest of the internet to notice.

Try it yourself

Live data on the SACS cable page. For context on Atlantic cables, see Equiano (Portugal ↔ South Africa, 2.5× floor), Marea (Virginia Beach ↔ Bilbao, 1.95× floor), and AMX-1 (coastal backbone, eight-country Latin American architecture).