Vietnam's cable faults test the redundancy bet

Four of Vietnam’s six international subsea cables are down at once, cutting roughly 30 percent of the country’s international internet traffic; Data Center Dynamics, citing VN Express, reports faults on the SJC2, AAE-1, AAG, and IA systems. Operators VNPT and Viettel are reporting network disruption, and Viettel has added 800Gbps of capacity on the IA cable and another 300Gbps through the Asia Pacific Gateway.

The SJC2 cable, which connects Vietnam to other parts of Asia, developed a fault last week in its S4 segment near Chung Hom Kok, Hong Kong, according to VN Express, and AAE-1 has had issues since February. The remaining two damaged systems — the Asia-America Gateway, or AAG, which links Southeast Asia with the US, and the Intra Asia cable, known as IA — are expected to begin repairs this month, though the specifics of the faults have not been disclosed.

Of Vietnam’s six subsea systems — AAG, IA, AAE-1, APG, ADC, and SJC2 (the last two being the newest) — four now carry faults. This is the third multi-cable failure since 2023: all five cables then connecting the country went down simultaneously that year, and in 2024 three cables — IA, APG, and AAE-1 — were hit by outages.

For investors, Vietnam’s experience is a case study in how cable redundancy gets mispriced: six cables look like diversification on a network map, but simultaneous faults are precisely what a diversified system is supposed to prevent. The repeated episodes suggest the shared points of failure are not inside any single cable but in the environment around them — landing stations, repair vessels, and the region’s limited construction capacity — and the SJC2 fault near Hong Kong is a reminder that a fault in one cable often lives close to a fault in another.

A cable count is not a resilience metric. For those underwriting subsea assets, what matters is independent restorability: physically separate landing points, terrestrial cross-border fallback, and repair contracts that are actually testable. Assets that can prove those characteristics should earn infrastructure pricing; assets that cannot are merchant risk wearing a map of the ocean floor.

This publication has argued that hyperscaler-anchored assets get infrastructure pricing while everything else fights for capital; the Vietnamese outage is the merchant end of that hierarchy in physical form. The traffic on AAG and IA is general international internet traffic, not a hyperscaler order book; when the cables fail, the revenue at risk is the operator’s own service revenue, and the mitigation is buying spot capacity in the middle of the incident. That is a different risk from a trans-Pacific cable with a take-or-pay contract behind it, and it should be priced differently.

Repairs are scheduled, and Viettel’s capacity purchases suggest the immediate damage can be managed. The harder question for any investor in digital infrastructure is what would have happened if the repair backlog had been longer than a month; the next time a market adds a cable, count not the new span but how many of the country’s routes would survive a simultaneous fault. That number is the real asset.