Two oceans, three cuts, one pattern: the submarine cable network shows its seams

The cables are not being cut in a coordinated way. They are being cut in an uncoordinated way that nonetheless keeps happening, and the cumulative pattern is what now concerns operators and defence officials. Three separate major cuts struck the Red Sea corridor during 2025 — PEACE on 4 March, SMW4 and IMEWE on 6 September — disrupting internet traffic between Asia, the Middle East, and Europe. On 2 January 2026, the BCS East cable between Latvia and Lithuania was damaged, the latest in a run of Baltic incidents that the regional governments have publicly discussed as suspected but unconfirmed sabotage.

For most of the internet's history, submarine cables have been a problem operators described as tractable. There are enough of them; traffic reroutes; the system is redundant by design. Those claims are still broadly true, but the margin has narrowed. Cloudflare and the RIPE NCC have separately documented that the Baltic and Red Sea events did produce measurable latency spikes and local outages in markets including Pakistan, India, and the United Arab Emirates. The Red Sea cuts are now attributed by the International Cable Protection Committee predominantly to commercial shipping activity — anchors dragging through a corridor with extremely dense traffic. The Baltic cuts have a messier attribution, with national authorities in Finland, Sweden, and Estonia repeatedly naming specific vessels.

The specific chokepoint that matters is the Bab el-Mandeb approach and the Gulf of Suez, where roughly sixteen cables converge in a narrow shipping lane now also transited by Houthi maritime activity. Repair in the Red Sea is slow under any conditions; it is slower when the nearest cable ships must route around an active conflict. The PEACE repair alone took months. Operators have quietly begun to price in the longer outages when marketing new routes.

The winners are alternative terrestrial backhaul providers, the Africa-coast cable systems that bypass the Red Sea corridor entirely, and the satellite-broadband operators whose disaster-recovery pitch now writes itself. The losers are the carriers whose service-level agreements were designed against a threat model in which Baltic and Red Sea cable cuts were independent, rare events rather than a correlated time series.

What the sequence of cuts forecloses is the comfortable fiction that the submarine network is a neutral piece of plumbing. It is a military object, an economic object, and a political object, all at once, and the diversification strategies to harden it — more routes, more cable ships, more spare capacity — are long-cycle capital projects that have not yet been funded at the scale the risk picture now suggests.