On 10 July 2026 the International Advisory Body on Submarine Cable Resilience approved its final report in Geneva. The body was set up by the ITU and the International Cable Protection Committee (ICPC), and its report is aimed at governments, regulators and industry. It is a good moment to look at what public sources actually say about cable cuts, without the drama that usually surrounds the topic.
This post does not predict an outage. It describes how cables fail, how long fixes take, and what a company can do about its own dependence on a handful of routes.
What the sources say
ENISA, the EU cybersecurity agency, published a news item on 31 August 2023 about its report on subsea cables. It states that "more than 97% of the world's internet traffic passes through subsea cables at some point." That is why the topic matters for any business with customers or staff in more than one country.
On causes, ENISA cites the ICPC's 2022 report: most incidents are accidental, mainly from anchoring and fishing. The ICPC's own page on dragged anchors (updated 24 February 2025) says anchors account for "approximately 30% of incidents each year representing around 60 faults."
ENISA also names the weak points. Cable landing stations and "chokepoints, where many cables are installed close to each other, are single points of failure." It treats a coordinated sabotage attack on several cables at once as the main malicious concern, and it says information on resilience, redundancy and capacity is lacking.
What is evidence and what is not
Evidence: the share of traffic carried by cables, the accidental nature of most faults, and the existence of chokepoints. These come from ENISA and the ICPC.
Not established by these sources: any claim that a given cut was deliberate, or that a coordinated attack is coming. ENISA says sabotage must be considered. It does not say it is happening or likely on a given date.
Repair is the slow part
ENISA notes that repairing a subsea cable is complex and needs specialised ships, of which there are few. The ITU advisory body's report is organised around three themes: faster deployment and repair, risk identification and mitigation, and greater geographic diversity of connectivity. The ITU page does not list figures, so we do not quote any.
For a business, the practical reading is simple. A cut is usually repaired, but not always in hours. Your design should assume some days of degraded paths in a region.
What to do now
- List the countries and regions where your customers, staff and cloud regions sit. Mark any that depend on one or two cable routes.
- Ask your network and cloud providers which diverse paths they use into those regions. Ask in writing and file the answer.
- Use more than one region for anything customer facing, and test failover to it. Higher latency is the likely failure mode, so set sensible timeouts and retries.
- Keep a small set of tasks that work with poor connectivity: cached documentation, offline runbooks, and a status page hosted away from your main provider.
- Review third party services that sit in a single region. Their cable exposure is your exposure.
How to check your own exposure
A traceroute from your office or build servers to your main endpoints shows how many network hops cross borders. It is crude, but it gives you a starting list.
traceroute api.example.com
Record the result next to your dependency inventory so the two stay linked.
Why it matters
Internet resilience is partly a physical topic. Software teams tend to think in availability zones and forget the seabed. Writing down which regions depend on which routes costs an afternoon and turns a vague worry into a risk register entry.
Quick check
According to the ICPC, what share of cable incidents come from dragged anchors?
About 30 percent, or around 60 faults a year, per its page updated 24 February 2025.
Does any cited source say a coordinated attack on cables will happen?
No. ENISA says it must be considered as a threat and names it as the main malicious concern, but it does not predict it.
