🛰️ 台灣灰色地帶與海底電纜監測

灰色地帶監測 深度文章

全球海纜產業
與斷纜風險

The Global Submarine
Cable Industry

海洋法專區 05 • 誰擁有海纜 • 為何斷 • 如何修

Law of the Sea 05 • Who Owns It • Why It Breaks • How It's Repaired

本頁重點

In short

全球約有 400 多條海底電纜、總長逾 110 萬公里，承載超過 95% 的國際資料流量。所有權正從電信業者組成的聯盟，轉向 Google、Meta、Microsoft、Amazon 等 OTT 巨頭的私有電纜。斷纜絕大多數是人為意外——漁業約 38%、拋錨約 25%（人為合計逾七成），環境因素約 14%；蓄意破壞佔比雖低卻最難歸責。修復受限於全球僅約 60 艘修纜船，常需 1–1.5 個月。把台灣放進這張全球版圖，會更清楚它為何同時是樞紐、也是前線。

There are about 400-plus submarine cables worldwide, totaling over 1.1 million km, carrying more than 95% of international data. Ownership is shifting from telecom consortia to private cables built by OTT giants — Google, Meta, Microsoft, Amazon. Most faults are accidental: fishing ~38%, anchors ~25% (over 70% human), with environmental causes ~14%; deliberate damage is a small share but the hardest to attribute. Repair is constrained by only ~60 cable ships worldwide and typically takes 1-1.5 months. Placing Taiwan on this global map shows why it is at once a hub and a front line.

海洋法專區 SERIES → 海洋權益Maritime Rights 國際法保護Int'l Law 執法管轄Enforcement 台灣海纜現況TW Cables 全球海纜產業Global Cables
此刻，在地球的海床上，超過 110 萬公里的玻璃絲正以光速傳遞著人類的對話、金流與記憶。它們大多只有水管粗細，卻撐起了整個數位世界。

這是一個低調到幾乎隱形、卻價值連城的產業。要看懂台灣的處境，得先看懂這張全球版圖。

一個隱形、卻撐起世界的產業

海底電纜產業很少出現在新聞頭條，卻是全球化最底層的硬體。它的規模、所有權結構與運作邏輯，過去十年正在劇烈改變——而這些改變，直接影響台灣這類關鍵節點的安全處境。

At this moment, on the planet's seabed, more than 1.1 million kilometers of glass thread are carrying humanity's conversations, money, and memories at the speed of light. Most are no thicker than a garden hose, yet they hold up the entire digital world.

It is an industry so understated as to be nearly invisible — and immensely valuable. To understand Taiwan's situation, first understand this global map.

An Invisible Industry That Holds Up the World

The submarine cable industry rarely makes headlines, yet it is the bottom-layer hardware of globalization. Its scale, ownership structure, and operating logic have changed dramatically over the past decade — and those changes directly shape the security situation of critical nodes like Taiwan.

規模：數字裡的全球海纜

400+全球海纜系統（運作＋規劃）
110萬總長度（公里，約）
95%國際資料流量倚賴海纜
~60全球專業修纜船

所有權的大轉向：從聯盟到 OTT 私有電纜

海纜的「誰出錢、誰擁有」過去十年發生了結構性轉變。

傳統模式 · 聯盟電纜 Consortium
多家電信業者共同投資、共享容量
傳統上，一條跨洋海纜由數家電信營運商組成聯盟，分攤建置成本、按比例分配容量。這種模式分散了風險，但決策慢、容量擴充受限。
新模式 · 私有電纜 Private Cable
OTT 巨頭獨資或主導興建
Google、Meta、Microsoft、Amazon 等內容與雲端巨頭（OTT），為了滿足自家資料中心之間爆量的內部流量，開始大舉投資、甚至獨資興建海纜（如 JUPITER、FASTER、PLCN 等系統）。海纜所有權正從電信業者，快速移向科技巨頭。
潛在容量 vs 點亮容量一條海纜舖設時的「潛在容量」（design capacity）遠大於實際「點亮」（lit capacity）使用的部分。業者會隨需求逐步「點亮」更多容量。這代表海纜的瓶頸往往不在玻璃本身，而在兩端的設備與商業決策。

Scale: The Global Cable Map in Numbers

400+Global cable systems (live + planned)
1.1MTotal length (km, approx.)
95%Of international data rides cables
~60Specialized cable-repair ships

The Great Ownership Shift: From Consortia to OTT Private Cables

"Who pays and who owns" has shifted structurally over the past decade.

Traditional · Consortium cable
Multiple telecoms co-invest and share capacity
Traditionally a trans-ocean cable was built by a consortium of telecom operators sharing construction costs and capacity pro rata. This spread risk but made decisions slow and capacity expansion limited.
New · Private cable
OTT giants build solo or lead
Content and cloud giants (OTTs) — Google, Meta, Microsoft, Amazon — began heavily investing in and even solely building cables (systems like JUPITER, FASTER, PLCN) to feed the surging internal traffic between their data centers. Cable ownership is moving fast from telecoms to tech giants.
Design Capacity vs Lit CapacityA cable's "design capacity" when laid far exceeds the portion actually "lit" in use. Operators progressively "light" more capacity as demand grows. So a cable's bottleneck is often not the glass itself but the equipment at each end and commercial decisions.

為什麼海纜會斷？人為佔了七成以上

很多人以為海纜斷裂多半是地震、海嘯等天災，但統計恰好相反：絕大多數斷纜是人為意外。

🎣
漁業活動（約 38%）底拖網、漁具勾掛是斷纜最大宗。漁船在淺水區作業，正好與海纜的脆弱上岸段重疊。
⚓
船舶拋錨（約 25%）大型船舶在不當區域下錨、或走錨拖行，會把海床上的纜線一併扯斷。這也是灰色地帶手法最常「假借」的姿態。
🌊
自然與環境（約 14%）海底地震、濁流、洋流長期磨損、地質滑動等。深海段相對安全，淺水與陸棚邊緣風險較高。
🕵️
蓄意破壞（佔比低、最難歸責）數量不多，卻是灰色地帶威脅的核心。它常被刻意偽裝成「漁業」或「拋錨」意外——關掉 AIS、在深夜或深水下手，讓事後幾乎無從區分意外與蓄意。
關鍵啟示人為意外佔大宗，反而成了蓄意破壞最好的「保護色」。當七成斷纜都是漁業與拋錨，一次精心偽裝的破壞要混進這個背景噪音，並不困難。這正是本站把行為訊號（走暗、異常徘徊、拖錨樣態）公開、供人工研判的理由。

Why Do Cables Break? Over 70% Is Human

Many assume cable breaks are mostly earthquakes and tsunamis, but the statistics say the opposite: the vast majority are human accidents.

🎣
Fishing activity (~38%)Bottom trawling and snagged gear are the single biggest cause. Fishing boats work the shallows, which overlap exactly with a cable's fragile landfall segments.
⚓
Ship anchors (~25%)Large ships anchoring in the wrong place, or dragging a dragging anchor, tear cables off the seabed. This is also the posture gray-zone tactics most often "borrow."
🌊
Natural & environmental (~14%)Seabed earthquakes, turbidity currents, long-term abrasion, geological slides. Deep-sea segments are relatively safe; shallow water and the shelf edge carry higher risk.
🕵️
Deliberate sabotage (small share, hardest to attribute)Few in number but the core of gray-zone threats. It is often disguised as a "fishing" or "anchor" accident — AIS off, acting at night or in deep water — making it nearly impossible to separate accident from intent after the fact.
The Key TakeawayBecause human accidents dominate, they become the best camouflage for deliberate damage. When 70% of breaks are fishing and anchors, a well-disguised sabotage can blend into that background noise. That is why this site publishes behavioral signals (going dark, anomalous loitering, anchor-drag patterns) for human assessment.

修復：為什麼要等上一個多月？

海纜一斷，修復不是「派人去接一下」那麼簡單。它是一場與深海、天氣與排程賽跑的工程。

修復為何耗時① 全球專業修纜船僅約 60 艘，分散在各區域，常需排隊調度；② 須先用儀器精確定位斷點；③ 修纜船到場後，用抓鉤撈起斷纜兩端、接續、測試，深水作業尤其耗時；④ 還得避開惡劣海象。綜合下來，通常要 1 至 1.5 個月。

修復時間越長，少數幾條海纜的節點（特別是外島）就越脆弱——這也呼應了上一篇談到的台灣外島困境。

防護與地緣：AIS、海纜布局與「海底長城」

📡
AIS 與海域感知業者與監測者透過 AIS 等船舶資料，監看海纜路線附近的異常船舶活動（低速徘徊、走暗、拋錨樣態），作為早期預警。本站正是這類開源監測的一環。
🧭
路線規劃與冗餘新建海纜會盡量分散路線、避開高風險漁區與拋錨區，並增加節點冗餘，降低「一斷全斷」的風險。
🌐
地緣布局海纜不只是商業設施，也是地緣戰略資產。中國透過國家級海纜布局（外界稱「海底長城」一類概念）與「一帶一路」的數位基礎設施輸出，擴大其在全球海纜版圖的影響力——這也讓海纜安全與國際政治更緊密交織。

Repair: Why the Wait of a Month or More?

When a cable breaks, repair is not "send someone to splice it." It is an engineering race against the deep sea, the weather, and scheduling.

Why Repair Takes So Long① Only ~60 specialized cable ships exist worldwide, spread across regions and often queued; ② the break must first be located precisely with instruments; ③ on arrival the ship grapples up both ends, splices, and tests — deep water is especially slow; ④ bad weather must be avoided. Together, it usually takes 1 to 1.5 months.

The longer the repair, the more fragile the nodes served by only a few cables (especially outer islands) — echoing the Taiwan outer-island dilemma from the previous article.

Protection & Geopolitics: AIS, Cable Routing, and the "Underwater Great Wall"

📡
AIS & maritime awarenessOperators and monitors use vessel data such as AIS to watch for anomalous activity near cable routes (low-speed loitering, going dark, anchor-drag patterns) as early warning. This site is one piece of that open-source monitoring.
🧭
Routing & redundancyNew cables route to diversify paths, avoid high-risk fishing and anchorage zones, and add node redundancy to reduce "one break, all down" risk.
🌐
Geopolitical positioningCables are not just commercial assets but geostrategic ones. Through state-level cable buildouts (concepts some call an "underwater great wall") and Belt-and-Road digital-infrastructure exports, China expands its weight on the global cable map — further entwining cable security with international politics.

常見問題 FAQ

全世界有多少條海底電纜？
依公開海纜地圖統計，全球運作中與規劃中的海底電纜系統約有 400 多條，總長度超過 110 萬公里，足以繞地球數十圈。這些海纜承載了全球超過 95% 的國際資料流量，是名副其實的網際網路骨幹。數字會隨新建與退役持續變動。
誰擁有這些海底電纜？
傳統上海纜由多家電信業者組成「聯盟」（consortium）共同投資、共享容量。但近年最大的轉變是 OTT 內容巨頭——Google、Meta、Microsoft、Amazon 等——大舉投資甚至獨資興建「私有電纜」，以滿足自家雲端與資料中心之間的龐大流量需求。海纜的所有權版圖，正從電信業者轉向科技巨頭。
海底電纜最常因為什麼而斷？
統計上，絕大多數斷纜屬於人為意外：漁業活動（拖網）約佔 38%、船舶拋錨約佔 25%，合計人為因素約佔七成以上；自然與環境因素（海底地震、洋流磨損等）約佔 14%。蓄意破壞雖然佔比不高，卻最難歸責，也正是灰色地帶威脅最棘手之處。
海底電纜斷了要多久才能修好？
通常需要 1 至 1.5 個月，甚至更久。原因之一是全球專業修纜船僅約 60 艘，常需排隊調度；加上要先精確定位斷點、再打撈、接續、測試，深水作業更耗時。修復時間長，正是少數幾條海纜的外島最怕斷纜的原因。

FAQ

How many submarine cables are there worldwide?
By public cable-map counts, there are roughly 400-plus live and planned submarine cable systems globally, totaling over 1.1 million km — enough to circle the Earth dozens of times. They carry over 95% of international data traffic and are the true backbone of the internet. The figures change continuously as systems are built and retired.
Who owns these cables?
Traditionally cables were built by consortia of telecom operators co-investing and sharing capacity. The biggest recent shift is OTT content giants — Google, Meta, Microsoft, Amazon — heavily investing in and even solely building "private cables" to feed the huge traffic between their cloud and data centers. Cable ownership is moving from telecoms to tech giants.
What most commonly breaks a submarine cable?
Statistically the vast majority are human accidents: fishing activity (trawling) ~38% and ship anchors ~25%, together over 70% human; natural and environmental causes (seabed quakes, current abrasion) ~14%. Deliberate sabotage is a small share but the hardest to attribute — the thorniest part of gray-zone threats.
How long does it take to repair a broken cable?
Usually 1 to 1.5 months, sometimes longer. One reason is that only ~60 specialized cable-repair ships exist worldwide and are often queued; plus the break must be located precisely, then grappled up, spliced, and tested — deep water is especially slow. Long repair times are exactly why outer islands served by only a few cables fear a break most.
← 04 台灣海纜現況 ← 04 TW Cables 返回深度文章 → Back to articles →

參考資料 Sources

本頁概念可對照以下公開、權威來源核實：

本文全球海纜規模、所有權轉向與斷纜成因比例，參酌 Vincent Chen-WS〈台灣海纜概況及備援機制〉（Medium, 2018）、TeleGeography 與 ICPC 公開資料。

本頁為研究與教育用途；統計數據為公開資料彙整，比例與條數可能隨年度與來源而異。詳見 方法論。

Sources

The concepts on this page can be verified against these public, authoritative sources:

The global scale, ownership shift, and fault-cause breakdown draw on Vincent Chen-WS, "Overview of Taiwan's Submarine Cables and Backup Mechanisms" (Medium, 2018), TeleGeography, and ICPC public data.

This page is for research and education; statistics aggregate public data, and proportions and counts can vary by year and source. See the methodology.

相關灰色地帶主題

Related gray-zone topics