Calculator Cables Locations Health Research
← All articles
Region

September 6, 2025: The Day the Red Sea Lost Its Cables — And the Internet Survived

At 05:45 UTC, the Alarms Started

On Saturday morning, September 6, 2025, network engineers across the Middle East woke up to a cascade of alerts. Something had gone very wrong beneath the Red Sea.

By 06:00 UTC, Microsoft had posted a terse notice on the Azure status page: "Network traffic traversing through the Middle East may experience increased latency due to undersea fiber cuts in the Red Sea." Within hours, the scope became clear — this wasn't a single cable fault. Multiple systems had been severed simultaneously near Jeddah, Saudi Arabia.

The affected cables read like a who's who of global internet infrastructure:

  • SEA-ME-WE-4 (Southeast Asia–Middle East–Western Europe) — one of the oldest and most heavily used Asia-Europe links
  • IMEWE (India–Middle East–Western Europe) — critical for Indian subcontinent connectivity
  • FALCON GCX — the Gulf backbone connecting Kuwait, Bahrain, Qatar, UAE, and Oman
  • Europe India Gateway (EIG) — a key alternative path between Mumbai and Marseille

Doug Madory, Kentik's director of Internet analysis, added a fifth: the AAE-1 (Asia-Africa-Europe-1) system showed degradation by September 9.

The Chokepoint Problem

To understand why this was so devastating, you need to understand the geography. The Red Sea narrows at the Bab el-Mandeb Strait — a corridor barely 30 kilometers wide between Yemen and Djibouti. Roughly 15 submarine cables thread through this passage, carrying an estimated 17–25% of all internet traffic between Asia, Europe, and the Middle East.

These cables don't spread out across the seabed. They run in parallel, bundled together in shallow waters where they're most vulnerable. An anchor dragged across the right stretch of seabed can — and did — sever multiple systems at once.

The International Cable Protection Committee (ICPC) attributes roughly 70% of all cable faults worldwide to anchors and fishing equipment. Their early analysis of the September 6 incident pointed to "commercial shipping activity" — likely a vessel dropping and dragging its anchor.

Who Felt It

The impact was immediate and widespread:

India and Pakistan: Pakistan Telecommunications Company Ltd (PTCL) confirmed reduced capacity and activated backup routes. Indian ISPs reported degraded international connectivity, particularly for traffic routed through Jeddah.

UAE: Customers of both du and Etisalat (e&) flooded social media with complaints. Both providers issued identical advisories acknowledging "slowness in data services due to an international submarine cable cut."

Microsoft Azure: The world's second-largest cloud platform confirmed increased latency for all traffic transiting the Middle East. Their engineering teams began emergency rerouting within minutes.

Kuwait: Authorities reported direct impact on the FALCON cable, which serves as the Gulf region's primary backbone.

NetBlocks, the internet monitoring organization, traced disruptions across at least ten countries.

How the Internet Survived

Here's what's remarkable: the internet didn't go down. It slowed, it degraded, some routes added 100-200ms of latency — but traffic kept flowing.

This is the Border Gateway Protocol (BGP) doing exactly what it was designed to do. When a path fails, routers automatically recalculate and redirect traffic through alternative routes. In practice, this meant:

  • Traffic from India to Europe that normally went straight through the Red Sea was rerouted eastward — through Singapore, across the Pacific to the US, then transatlantic to Europe. A path that normally takes ~90ms suddenly took 250-300ms.
  • Some providers shifted traffic to terrestrial routes across Central Asia — through Iran, Turkey, and into European networks.
  • Content Delivery Networks like Cloudflare and Akamai served cached content from local nodes, masking much of the disruption for end users.

ThousandEyes data showed that while latency spiked significantly on routes like Mumbai-Frankfurt, packet loss remained negligible for most monitored connections. The internet was slower, but it was alive.

The Repair Challenge

Fixing submarine cables in the Red Sea is not like fixing a cable anywhere else. The region has been a conflict zone since late 2023, with Yemen's Houthi forces attacking commercial shipping. This creates three compounding problems:

1. Permits: Cable repair ships need transit permits from multiple nations, including parties to the conflict. In previous incidents, these permits took weeks to obtain.
2. Safety: Repair vessels are large, slow-moving targets in a zone where commercial ships have been hit by missiles and drones.
3. Global repair ship shortage: As of 2025, the world has fewer than 60 cable repair vessels. Many are already committed to new cable installations or other repair jobs. The queue is months long.

For the February 2024 Red Sea cuts (AAE-1, SEACOM, and EIG), repairs took until July 2024 — five months. The September 2025 cuts faced similar timelines.

What GeoCables Sees Today

We monitor all major Red Sea cables from our four probe servers in Minsk, Almaty, Tbilisi, and Jerusalem. Here's what our latest measurements show:

CableTargetRTT (ms)Status
SEA-ME-WE-4Marseille237Normal
SEA-ME-WE-5Toulon237Normal
FALCONSuez255Normal
EIGMumbai180Normal
AAE-1Marseille223Normal
SEACOMMtunzini296Normal

All six cable systems that transit the Red Sea are currently showing normal RTT values from our probes. The repairs are complete, traffic has returned to optimal paths, and latency has normalized.

But "normal" is deceptive. These cables will be cut again. The Bab el-Mandeb Strait remains a conflict zone, anchors will continue to drag, and the fundamental vulnerability — 15 cables in a 30km corridor — hasn't changed.

The Deeper Question

The September 2025 Red Sea cuts weren't unprecedented — similar incidents happened in February 2024 and March 2025. What's changed is the frequency. The Red Sea has gone from a theoretical chokepoint to a proven, recurring failure mode for the global internet.

The industry's response is already underway. New cable routes are being planned that avoid the Red Sea entirely — through the Arctic, across Central Asia, or around the Cape of Good Hope. But these alternatives add thousands of kilometers and tens of milliseconds of latency.

For now, the Red Sea remains what it has been for a century: the shortest path between East and West. And as long as that's true, the world's internet will keep running through a 30-kilometer bottleneck in one of the most dangerous stretches of water on Earth.


GeoCables monitors all major Red Sea cable systems 24/7 from probes in Minsk, Almaty, Tbilisi, and Jerusalem. View current status →

Evgeny K.
Written by
Evgeny K.
Infrastructure Engineer · Founder of GeoCables
Built GeoCables to monitor submarine cables in real time. Runs a private network of 4 measurement servers with RIPE Atlas probes in Minsk, Almaty, Tbilisi, and Jerusalem.

🌐 Log In

Access your routes, favorites, and API key

Create account Forgot password?