Author: CryptoSlate
Compiled by Deep潮 TechFlow
DeepOcean Summary: A Cambridge University research team, using 11 years of data and 68 submarine cable outage events, demonstrated that cutting undersea cables has almost no impact on the Bitcoin network. However, they identified a genuine vulnerability—not underwater, but at cloud service providers such as Hetzner, AWS, and Google Cloud. The study’s findings strongly refute claims of Bitcoin’s fragility while providing a quantitative framework for identifying real infrastructure risks.
The full text is as follows:
In March 2024, underwater disturbances off the coast of Côte d'Ivoire severed seven submarine cables, causing a regional internet disruption with an IODA score exceeding 11,000.
For Bitcoin, the global impact is negligible. Only about five nodes were affected, accounting for approximately 0.03% of the entire network, with an impact magnitude of -2.5%, well within normal fluctuation ranges.
No price volatility, no consensus disruption.
A new Cambridge study, covering 11 years of Bitcoin network data and 68 verified undersea cable failures, concludes that undersea cable disruptions have historically had extremely limited impact on the Bitcoin network.
In comparison, coordinated attacks against a small number of custodial networks are an order of magnitude more effective at disrupting visible nodes than random infrastructure failures.
Notably, China’s mining crackdown and the global adoption of censorship-resistant infrastructure may inadvertently be pushing Bitcoin toward a more resilient network topology.
Tor has long been regarded as a privacy tool and is now becoming a structural resilience layer, with the majority of Bitcoin nodes running on Tor.
Empirical data contradicts the concerns.
Cambridge University researchers Wenbin Wu and Alexander Neumueller constructed a dataset covering 2014 to 2025, including 8 million Bitcoin node observations, 658 submarine cables, and 385 cable outage events, cross-referenced with interruption characteristics.
Of the 385 reports, 68 matched verifiable outages, with 87% of validated cable events resulting in node changes below 5%. The average impact was -1.5%, with a median of -0.4%.
The correlation between node outages and Bitcoin price is nearly zero (r = -0.02). Cable failures that dominate regional headlines typically leave no trace in Bitcoin’s decentralized network.
The study models Bitcoin as a multilayer network: a physical connectivity layer of 225 countries connected by 354 undersea cable links, a routing infrastructure layer (autonomous systems), and the Bitcoin peer-to-peer overlay layer.
Under random cable removal, the critical threshold for more than 10% of nodes becoming disconnected lies between 0.72 and 0.92. Before Bitcoin experiences meaningful fragmentation, nearly all international cables must fail.
Where are the real vulnerabilities?
Targeted attacks operate very differently. Randomly removing cables requires removing 72% to 92% of cables to reach the threshold of 10% node disconnections; for targeted attacks on cables with high betweenness centrality, this proportion drops to 20%.
The most effective strategy is to target top autonomous systems (ASNs) by the number of nodes, requiring only the removal of 5% of routing capacity to reach the threshold.
The author characterizes this ASN targeting scenario as "hosting provider shutdowns or coordinated regulatory actions, rather than actual physical cable cuts." The top networks identified by the model include: Hetzner, OVHcloud, Comcast, Amazon Web Services (AWS), and Google Cloud.
A March 2026 Bitnodes snapshot confirms this distribution: among 23,150 accessible nodes, Hetzner hosts 869, Comcast and OVH each host 348, Amazon hosts 336, and Google hosts 313.
This is not an assertion that "five suppliers can destroy Bitcoin."
Even if the public network were completely removed, most nodes would still operate because Tor carries the majority of the network. However, this finding reveals where coordinated actions could cause connection disruptions and propagation outages that random cable failures would not.
Recent cloud service outages illustrate this category of risk. Amazon attributed an outage in March 2026 to a failed software deployment, and other reports described an outage in AWS's Middle East region following an attack on its data center.
These events had no meaningful impact on Bitcoin, but they demonstrated that custodian-related failures are real, not just theoretical assumptions.
Tor as a structural resilience layer
The composition of the Bitcoin network has changed significantly.
Tor adoption grew from nearly zero in 2014 to 2,478 nodes (23%) in 2021, and then to 7,617 nodes (52%) in 2022. By March 2026, 14,602 out of 23,150 accessible nodes were Tor nodes, accounting for 63%. This growth coincided with multiple censorship events: Iran’s internet shutdown in 2019, Myanmar’s coup in 2021, and China’s mining ban in 2021.
Node operators independently shifted to censorship-resistant infrastructure, demonstrating the network's adaptive, self-organizing capability.
Tor presents a challenge: most Bitcoin nodes are now unobservable in terms of location.
The author addresses this issue by constructing a four-layer model that treats the Tor relay infrastructure as an independent network layer. Tor relays are physical servers with known locations.
Using consensus weight data from 9,793 relays, the authors modeled how cable failures disconnecting countries could simultaneously take relays offline.
The research findings were unexpected. The four-layer model consistently produced a higher critical failure threshold than models considering only the public network, with improvements ranging from 0.02 to 0.10.
The majority of Tor relay consensus weights are concentrated in Germany, France, and the Netherlands—countries with extensive cable connectivity. A cable outage disconnecting peripheral nations would not diminish the relay capacity of these well-connected countries.
Attackers must remove more infrastructure to simultaneously disrupt public network routing and Tor circuits.
China factor
Bitcoin's resilience hit its lowest point of 0.72 in 2021, coinciding with the peak of hash rate concentration.
Cambridge data shows that in 2019, 74% of hashing power was located in East Asia. The geographic concentration of nodes caused the public network's resilience to decline by 22% between 2018 and 2021.
The 2022 rebound was very strong. Following China's mining ban, as infrastructure became more decentralized, the threshold rose to 0.88, and Tor adoption accelerated in tandem.
Although the author avoids simplistic causal claims, regulatory pressure has driven geographic redistribution and the adoption of censorship-resistant infrastructure—both of which have enhanced network resilience.
The apparent centralization is partly due to measurement error. As Tor adoption increases, the public network sample becomes concentrated in fewer locations, causing the Herfindahl-Hirschman Index to rise from 166 to 4,163, while Hetzner’s actual share decreases from 10% to 3.6%. This concentration reflects a change in sample composition, not true centralization.
Cloud services are the real risk.
Concerns over the security of undersea cables will continue to intensify. Investigations in the Baltic Sea, the European Commission’s security toolbox, and reports on Russian infrastructure all point to ongoing geopolitical anxiety.
For Bitcoin, historical data shows that most cable events are noise.
The truly significant infrastructure issue is whether policy coordination, cloud service outages, or hosting restrictions can cause connectivity disruptions at the autonomous system level.
The operational threshold for ASN-directed scenarios is 5% of routing capacity, which represents the critical point at which publicly accessible network nodes experience significant disruption, not consensus failure.
Tor's majority share provides a baseline in extreme scenarios. Protocol-layer mechanisms not included in this study—such as block relay networks, compact block relay, and Blockstream Satellite—add additional layers of resilience, making the estimates conservative.
Bitcoin is not as fragile as critics imagine, but it is also not entirely disconnected from infrastructure.
The network demonstrated graceful degradation under pressure rather than catastrophic failure. Censorship pressure accelerated the adoption of infrastructure, which in turn strengthened resilience against coordinated risks.
The threat model centered on submarines cutting undersea cables overlooks closer bottlenecks: a handful of networks where coordinated actions can cause temporary disruptions without requiring dramatic undersea operations or acts of war.