The Bitcoin network has been running non-stop since 2009. The question of what it actually takes to break through this is one that no one has ever rigorously answered.
Researchers at the Cambridge Center for Alternative Finance last week published the first longitudinal study of the Bitcoin blockchain’s resilience to physical infrastructure disruptions, analyzing 11 years of peer-to-peer network data for 68 verified undersea cable failure events.
The headline finding is that between 72% and 92% of the world’s undersea cables between countries would have to fail simultaneously before Bitcoin would experience a significant node disconnection.
In a world where the Strait of Hormuz is currently disrupted and infrastructure vulnerabilities are in the spotlight, this study provides the first empirical benchmark of how difficult Bitcoin really is to be knocked offline.
This number suggests that the network will degrade gracefully, rather than collapsing catastrophically. The researchers ran 1,000 Monte Carlo simulations for each scenario across the dataset and found that very few random cable failures were recorded.
More than 87% of the 68 real-world cable failure events they investigated had less than 5% impact on the nodes. The single largest event was in March 2024 when an undersea disturbance off the coast of Ivory Coast damaged seven to eight cables simultaneously, shutting down 43% of regional nodes, but only affecting five to seven Bitcoin nodes globally, or about 0.03% of the network.
The correlation between cable failure and Bitcoin price was effectively zero, at -0.02. Infrastructure disruption is invisible to daily price fluctuations.

However, the most important finding of this paper is the asymmetry between random and targeted attacks.
Random cable failures require 72-92% removal to cause damage, while targeted attacks on cables with the highest inter-cable centrality, which act as intercontinental chokepoints, reduce that threshold to 20%.
Additionally, if you target the top five hosting providers by number of nodes: Hetzner, OVH, Comcast, Amazon, and Google Cloud, you only need to remove 5% of their routing capacity to achieve the same effect.
This is a fundamentally different threat model. Accidental failure is a natural phenomenon. A targeted attack is an act of a state, a coordinated outage due to regulation of a hosting provider, or the intentional disconnection of a critical cable route. This study essentially maps two very different adversaries. One is Bitcoin, which can easily survive, and the other is still a reliable risk.
How threats to Bitcoin change over time
This paper tracks how resilience has evolved over time, but the trajectory is not a straight line. Bitcoin was most resilient in its early years from 2014 to 2017, when the network was geographically diverse and the critical failure threshold was around 0.90 to 0.92.
As the network grew rapidly but became geographically concentrated, resilience declined sharply from 2018 to 2021, reaching a low point of 0.72 in 2021 at the peak of mine concentration in East Asia. China’s mining ban in 2021 forced redistribution, and resilience partially recovered to 0.88 in 2022, but settled to 0.78 in 2025.
The discovery of TOR challenges conventional thinking. As of 2025, 64% of Bitcoin nodes are using TOR, making their physical location unobservable.
It has been believed that this unobservability may hide vulnerabilities, and that if TOR nodes are found to be geographically concentrated, the network may be more vulnerable than it appears.
The Cambridge researchers built a four-layer model to test this and found the opposite. TOR relay infrastructure is concentrated in Germany, France, and the Netherlands, which have extensive submarine cables and land border connections.
These countries are among the most difficult to disconnect, so attackers face compounding problems when attempting to disrupt TOR relay capabilities by cutting cables. The four-layer model consistently showed higher resilience than the clearnet-only baseline, with TOR adding 0.02 to 0.10 to the critical failure threshold.

The paper calls this “adaptive self-organization.” In response to censorship incidents such as Iran’s internet shutdown in 2019, Myanmar’s coup d’état in 2021, and China’s mining ban, adoption of TOR has increased rapidly.
The Bitcoin community has moved to a censorship-resistant infrastructure without central coordination, and that move has made the network less susceptible to physical interruptions.
With the Strait of Hormuz effectively closed and regional wars disrupting infrastructure across the Middle East, the question of what will happen to Bitcoin if undersea cables are damaged is not theoretical.
This research suggests that the answer is probably nothing, unless someone intentionally targets the specific cable and hosting providers that matter most.

