Debate over satellite versus submarine internet

X users spent May 24 arguing over a familiar internet question: can satellite service really rival the undersea cables that carry most global traffic. The exchange was driven in part by post 2058495871541764507, which dismissed satellite performance in blunt terms and framed submarine cables as the only serious backbone for internet service. Other posts pushed back, pointing to the reach of low-Earth-orbit systems such as Starlink and the fact that satellites can serve places fiber and cable do not reach. The underlying facts are less dramatic than the social-media version. The International Telecommunication Union says submarine cables carry about 99% of the world’s internet traffic and remain the backbone of international communications. TeleGeography, a telecom research firm, says it can confirm that submarine cables account for more than 99% of intercontinental data traffic. ### If cables carry almost everything, why are satellites even in the debate? Satellite internet is in the debate because coverage and backbone capacity are different questions. ITU says submarine cables connect continents, markets and households, but it also says they are crucial for extending connectivity to underserved and remote areas. In practice, satellites matter most where terrestrial networks are costly, sparse or damaged. (itu.int) Starlink, the best-known low-Earth-orbit service, has also improved the case for satellite on everyday responsiveness. SpaceX said in a public latency note that U.S. median Starlink latency during peak hours fell by more than 30%, to 33 milliseconds from 48.5 milliseconds, and that worst-case peak-hour latency dropped to less than 65 milliseconds from more than 150 milliseconds. ### Why do cable supporters keep saying latency still favors fiber and subsea links? (itu.int) Latency remains the strongest argument for cable because data traveling over fiber follows a more direct and stable terrestrial or subsea path than data that must go up to a satellite and back down. SpaceX itself says Starlink’s round-trip delay includes propagation from the user to the satellite and back to ground, plus additional delay when traffic uses laser links or longer ground-network paths. (starlink.com) The U.S. Federal Communications Commission’s Measuring Broadband America program treats latency as one of the core performance measures for consumer broadband because it affects web browsing, calls and other real-time uses. The FCC report does not stage a simple satellite-versus-submarine comparison, but it underscores why users fixate on latency when they compare technologies. (starlink.com) ### Does that mean the “satellite is fake” line is wrong? Yes, if “fake” means unusable. SpaceX’s published figures show modern low-Earth-orbit satellite service can deliver latency low enough for video calls, web browsing and many interactive applications. That is very different from older geostationary satellite systems, which were defined by much higher delay. No, if the claim is that satellites have not replaced cables as the core of the global internet. (fcc.gov) ITU says submarine cables remain the backbone of global communications, and TeleGeography says more than 99% of intercontinental traffic still moves over them. Capacity and traffic share are not close. ### What about reliability when a cable breaks or a satellite network gets congested? (starlink.com) Reliability depends on the kind of failure. ITU says submarine cables are vulnerable mainly to fishing, anchoring and natural hazards such as earthquakes and underwater landslides, and it said more than 170 cable repairs were reported worldwide in 2025. Those failures matter because so much international traffic depends on cable routes. (itu.int) Satellite systems face different constraints. SpaceX says Starlink latency can worsen when traffic must be routed over laser links or when gateway and internet connection points are not close to the user’s destination. That means satellites can avoid some cable-specific risks, but they still depend on network design, spectrum and ground infrastructure. ### So what is the cleanest way to read the argument? (itu.int) The cleanest answer is that submarine cables and satellite internet do different jobs. ITU’s traffic figures show cables dominate global and intercontinental transport, while Starlink’s own performance data shows satellites have become much more competitive for end users who need coverage where wired options are weak or absent. (starlink.com) On May 24, the next place to check the argument was the same one that started it: X posts such as 2058495871541764507, alongside Starlink’s published latency note and ITU’s cable-resilience material. (starlink.com) (itu.int)