Big Tech Builds Undersea Cable Web as AI Escalates Data Demand

AI推高数据需求，科技巨头自建海底光缆网 像 Google 和 Meta 这样的超大规模科技企业目前出资建设了三分之二的新建海底光缆，以满足 AI 驱动下对数据的巨大渴求；尽管地缘政治风险不断升级，实体海底光缆仍成为现代科技基础设施不可或缺的支柱。 本期节目深入探讨了全球互联网鲜为人知的物理实体，揭示了超过95%的国际数据流量都依赖庞大的海底光缆网络进行传输，且其路线与19世纪的电报线路高度吻合。作家 Samanth Subramanian 概述了海缆所有权如何从国家支持的电信财团转向由 Google、Meta 等现代科技巨头直接掌控，同时解释了物理损坏、海洋地理以及地缘政治紧张局势如何对这一关键基础设施构成威胁。对话最终强调，随着 AI 的爆发和指数级增长的数据需求，在可预见的未来，人类仍将紧密依附于这些海底玻璃纤维细丝。 互联网是一项高度依赖实体设施的网络，其运转依托于一个与19世纪电报路线高度重合的全球海底光缆网络。 (0:00) 海底光纤利用波分复用（WDM）技术传输数据，通过细如发丝的玻璃纤维传输高速激光脉冲。 (8:00) 卫星和蜂窝基站无法承载全球海量的数据流量，这使得海底光缆对现代互联网而言不可替代。 (8:00) 海底光缆的所有权已从国家主导的电信财团转向大型科技巨头（Google、Meta、Microsoft、Amazon），目前大部分新建海缆均由它们出资建设。 (11:00) 科技巨头对海底互联链路的直接掌控，在数据隐私、安全和数字接入方面给发展中国家带来了严峻的主权困境。 (25:00)

Despite common assumptions that the internet operates wirelessly, global digital connectivity remains bound to an immense physical network of undersea fiber-optic cables, which are increasingly funded by tech hyperscalers and exposed to acute geopolitical tensions. Transition to Big Tech Infrastructure Dominance The financing of global subsea cables has undergone a fundamental transformation over the last four decades. In the 1980s and 1990s, state-owned national telecom monopolies formed consortiums to share bandwidth, followed by a wave of speculative private capital during the dot-com era. Today, the sector is dominated by four cash-rich tech giants—Google, Meta, Microsoft, and Amazon—which now own or fund, in whole or in part, two out of every three new cables. Author Samanth Subramanian notes that laying a modern transatlantic route costs roughly $500 million. Because hyperscalers generate and consume the bulk of international data traffic, deploying proprietary cables is more cost-effective than leasing bandwidth from third parties. Unlike railway monopolies, the global network of 500 to 550 active subsea cables relies heavily on redundancy, with tech firms routinely buying backup capacity on competing cables to prevent downtime. Physical Fragility and Victorian-Era Repair Methods (20:00) Undersea cables closely follow 19th-century telegraph paths and maritime trade routes, converging on major chokepoints like the Red Sea, the Suez Canal, and the Strait of Hormuz. In India, cables continue to land in Mumbai and Chennai at the exact sites selected over a century ago. This concentration creates structural vulnerability; an underwater volcanic eruption in 2022 severed the sole international cable connecting the island nation of Tonga, cutting off the archipelago's connectivity. While roughly 100 cable cuts occur globally each year—primarily caused by commercial fishing trawlers and dragged anchors—large-scale internet outages are avoided through automated traffic rerouting. When breaks occur, repair processes remain surprisingly low-tech: specialized ships drop grapnel hooks to drag the seafloor, snag the severed cable, and haul it aboard, where technicians splice hair-thin glass fibers in stabilized clean rooms. Geopolitical Fragmentation and the Threat of Bifurcation (30:00) Rising geopolitical conflict between the United States and China is actively reshaping undersea infrastructure. U.S. sanctions and security measures have dissolved historical consortia that brought Western firms and Chinese state-backed suppliers like HMN Tech together. Consequently, direct subsea links between the U.S. West Coast and China have halted, forcing data routes through intermediaries like the Philippines or Singapore. Industry observers warn this separation could permanently bifurcate the internet into competing, redundant systems. Governments are also treating cables as high-risk targets for sabotage and hybrid warfare. The UK now deploys two naval vessels to patrol its undersea lines, Scandinavian nations are monitoring Baltic activity, and Taiwan harbors concerns over its reliance on just 15 international cables. Concurrently, big tech's expansion raises sovereignty dilemmas: Meta's 2Africa cable brings connectivity to underserved regions, yet developing nations often lack the regulatory leverage of European states to impose domestic data storage requirements. Bandwidth Demands and the Accelerating AI Boom (39:30) Subramanian rejects the notion of a 'post-cable' wireless future. While satellite networks like Starlink can offer low-bandwidth communication or disaster relief—as demonstrated in Tonga—they cannot support the massive data transfer required by 4K video, cloud infrastructure, and enterprise databases. The expansion of artificial intelligence has intensified this dynamic. Centralized training workflows and real-time processing generate cross-border data flows that require the ultra-high throughput of optical fibers using wavelength-division multiplexing. Rather than causing an infrastructure glut, the AI boom is serving as a major demand catalyst, accelerating the pace of capital deployment by hyperscalers into new deep-sea routes. While the physical layout of subsea cables still mirrors Victorian telegraph maps, escalating geopolitical rivalries and surging artificial intelligence workloads have turned ocean-floor glass fibers into a high-stakes arena of state competition and private monopolization.

作家兼记者萨曼思·苏布拉马尼安（Samanth Subramanian）指出，看似无形的全球互联网与人工智能算力，本质上高度依托于极其脆弱的海底光纤网络；随着科技巨头掌控大部分海缆建设以及中美地缘博弈加剧，这一关键物理基础设施正面临所有权垄断、航路咽喉受阻与全球网络分裂的深刻演变。 物理底座与工程现实：光纤不可替代 现代互联网并非漂浮在空中的无线数字信号，而是深植于海底的物理光纤网络。深海光缆的核心是细如发丝的超纯玻璃纤维，依赖激光脉冲和波分复用（WDM）技术在单根纤芯上以不同光频率同时传输多路数据。尽管低轨卫星星座（如 Starlink）在汤加2022年火山喷发导致唯一国际海缆断裂后曾提供短期应急通信，但在处理4K流媒体、云计算及金融交易等海量跨境数据时，无线电波与卫星受制于带宽瓶颈，根本无法替代海缆。 这一底层物理网络的部署极为依赖少数几家垄断实体与精密工程。专用勘测船必须提前勘测避开海底悬崖、海沟与复杂洋流的最优路径，布缆船则需以极其严苛的速度释放滚筒光缆——速度过快会拉断脆弱光纤，过慢则会导致电缆在海床过度松弛。目前全球仅有少数专业厂商具备制造高规格深海光缆的能力，且几乎所有布缆船的航速与航迹测算导航软件均被全球唯一一家公司垄断，显露出极高的工程集中度与技术脆弱性。 投资范式转变：科技巨头的基建控制 过去四十年间，海底光缆的投融资模式历经了三次根本性洗牌：从20世纪80至90年代国有电信巨头分摊资本与带宽的“国家联盟模式”，到90至2000年代由私人开发商筹资并预售容量的“投机私有化模式”，再到当今以谷歌（Google）、Meta、微软（Microsoft）和亚马逊（Amazon）为主导的“科技巨头直营模式”。如今一条横跨大西洋的海缆成本高达约5亿美元，由于超大规模计算中心产生了全球绝大多数互联网流量，直接出资自建专属光缆比长期向第三方电信运营商租用带宽更具成本效益。 当前全球500至550条活跃海缆中，有三分之二的新建系统由这四家科技巨头全资或联合出资持有。与19世纪避免重复铺轨的铁路垄断不同，海缆生态极为讲求冗余，巨头在拥有自建线路的同时仍会向竞争对手购买备份容量以防单点故障。然而，这种资本垄断对欠发达国家构成了严峻的数字主权困境。例如Meta出资修建的环非洲“2Africa”超长海缆虽能弥补接入空白，但发展中国家话语权微弱；西欧国家可强制要求外企在本土设立数据中心，而像科特迪瓦这样的国家却面临在放弃数据隐私监管与被科技巨头完全绕过之间的残酷抉择。 地缘博弈、航运咽喉与网络分化风险 (30:00) 海底光缆的全球拓扑网络深受历史惯性与海洋地理的制约，其核心线路高度重合于19世纪大英帝国的电报航线和现代海运航道（例如印度的海缆登陆点至今仍集中在孟买与金奈）。红海、苏伊士运河、霍尔木兹海峡以及埃及等狭窄通道汇聚了亚欧之间极高密度的网络流量，这种地理集中性使全球数据流极易因自然灾害或恶意破坏而遭受大范围拥堵。全球每年发生约100起海缆物理切断事故，虽大部分源于商业拖网渔船或违规抛锚，但地缘冲突使得国家层面的蓄意切断与“灰色地带战争”威胁急剧攀升。 为应对潜在安全风险，英国已派遣两艘军舰常态化巡逻以保护水下通信与能源管线，波罗的海及北欧国家提高了警备等级，而仅依赖15条国际海缆的中国台湾也对其网络韧性高度担忧。更为深远的冲击来自中美博弈导致的网络割裂。过去西方企业与中国华海通信（HMN Tech）联合建缆的局面已被打断，在美国制裁限制下，跨越美中两国的直连海缆建设基本停滞，数据被迫经由菲律宾或新加坡迂回绕道。业内警告，若外交僵局持续，全球可能走向相互脱钩、标准割裂的双轨电缆体系，彻底颠覆统一万维网的初衷。 原始修复机制与AI引发的建设繁荣 (39:30) 相较于前沿的光学通信，深海光缆的维护维修工艺却呈现出极其原始的特征。当水下电缆被切断后，专业维修船仍主要沿用19世纪电报时期的传统手段——向故障海域投放沉重的铁制抓锚沿海床拖拽，勾住断缆拉上甲板后，由工程师在船载无尘实验室内手工熔接玻璃纤芯，再小心放回海床。此外，不同于跨国金融机构使用的陆上私有光纤，完全与公共互联网物理隔离的越洋级独立专线几乎专属于美中等大国军方，其海缆资产由军方雇佣民用机构敷设并受到严密保护。 与此同时，人工智能（AI）的大爆发正在为海底基建注入强劲的增长周期。以 ChatGPT 和 Claude 为代表的生成式大模型依赖分布式算力中心之间庞大的跨国数据同步，彻底消除了数年前市场关于海缆产能过剩的疑虑。海缆行业历来伴随宏观经济剧烈波动（如2000年代初互联网泡沫破灭后曾陷入长达五年的投资停滞），但访谈指出，当年电信泡沫时期的超前建设并非毫无意义的浪费，而只是预期与需求兑现之间的时间差；眼下爆发的AI算力浪潮正如同过去的电网扩容，促使各大科技巨头重金加速新一轮海底基础设施布局。 本期节目展现了现代数字文明底层的一个核心矛盾：承载最新AI革命与全球金融运转的关键命脉，在物理上仍高度依赖脆弱的玻璃纤维丝、受限的海洋地峡以及原始的抓锚修复工艺。尽管生成式AI的需求激增消除了过度投资的隐忧，但大国博弈引发的中美直连中断、航线改道以及国际公海基础设施防御的法律与军事盲区，意味着全球互联网在算力狂飙的同时，正步入地缘碎片化与物理脆弱性并存的动荡期。

## Background 本期音频由彭博（Bloomberg）旗下的《Odd Lots》栏目于 2026 年 5 月 13 日在播客及网络平台发布。根据官方简介，前参议员特德·史蒂文斯曾将互联网形象地称为“一系列管道”，而现实中穿行于大洋底部的光纤电缆正是维系全球网络运转的脆弱物理中枢；节目邀请了《海浪之下的网络》一书作者萨曼斯·苏布拉马尼安（Samanth Subramanian），深入探讨海底光缆系统背后奇特的历史矛盾、严苛的海洋地理瓶颈以及保护这些关键基础设施所面临的巨大挑战。 ## TL;DR 节目深入解构了现代互联网背后被普遍忽视的物理骨架，指出所谓轻盈的“云端”与 AI 实质上完全受制于深海中重资产、高脆弱性的光纤网络。内容系统梳理了海缆行业从 19 世纪电报线缆、国有电信联合体，演变至如今由少数科技巨头（Big Tech）资本主导铺设的格局转移，并详细展示了由单一垄断软件控制的铺设航行与充满时代错位感的海底捞缆抢修工艺。同时，讨论重点剖析了红海与霍尔木兹海峡等地理咽喉面临的切断风险、中美地缘竞争催生的网络分叉隐忧，以及军用物理隔离网络的现实。最后，嘉宾与主持人驳斥了低轨卫星能够替代海缆的虚假幻想，指出生成式 AI 爆发带来的数据饥渴只会进一步加速海底光缆的大规模扩张周期。 ## Atmosphere & style 整期对话充满知性探索欲与轻松幽默的思辨张力，主持人默契的抛梗接梗让深奥冷门的海洋工程与宏观地缘议题变得引人入胜。听觉体验带有一种强烈的认知颠覆感——将听众从对数字世界轻灵无形的错觉中，一把拽入数千米冰冷大洋底部与生锈抓钩、巨型卷轴以及发丝般纤细的玻璃纤维面对面。在对大洋深处光信号全反射这一人类工程奇迹由衷赞叹的同时，基建巨头垄断、灰色地带战争以及数字主权博弈的现实又为节目平添了一层冷静而发人深省的警示色彩。 ## Bulletin - `13:16` 全球约三分之二的新建海底光缆现由 Google、Meta、Amazon 和 Microsoft 四大科技巨头部分或全资拥有，建设主导权已从早期的国有电信联合体彻底转移至民营科技巨头手中。 - `20:00` 全球大洋底部运行着约 500 至 550 条海缆，每年发生约 100 起切断事故，主要源于船锚拖曳与渔船深海拖网，但依托密集的冗余网络与多厂商互保机制，日常流量几乎无感切换。 - `30:00` 地缘政治制裁导致中美直连海缆多年停摆，跨太平洋海量数据被迫绕行菲律宾或新加坡，甚至引发因互不信任而在同一区域重复建设两套海缆的“系统分叉”现象。 - `38:15` 针对卫星替代海缆的技术幻想，业内共识是低轨卫星带宽无法承载 AI 爆发与流媒体带来的工业级数据体量，海缆仍将承担全球 95% 以上的跨洲数据主干传输。 ## Key takeaways - `00:00` 数字化世界并非轻盈漂浮的虚拟存在，其底层完全依赖高耗电数据中心与横跨大洋海床的脆弱物理光纤网络。 - `05:00` 2022 年汤加火山喷发引发海底滑坡直接切断该国唯一的对外光缆，令其陷入数字黑暗，暴露出全球对单条物理缆线的极端脆弱性。 - `07:20` 现代海缆采用波分复用（WDM）技术在头发丝般细的高纯度玻璃中并发传输多频段激光，但全球制造能力与铺缆导航控制软件却被极少数供应商高度垄断。 - `11:47` 由于卫星信道存在难以突破的物理带宽上限，全人类每天消费的海量流媒体、云端算力与高频金融数据只能依赖深海“管道”流转。 - `13:16` 海缆出资模式历经国有电信联合体、私人投资驱动到 Big Tech 巨头自建三个阶段，单条跨大西洋光缆造价约 5 亿美元对万亿级巨头而言极易负担。 - `17:42` 与铁路的“天然垄断”不同，海缆极度依赖冗余容灾，巨头之间普遍存在反向采购对方备用容量的互保竞合网络，以实现零停机。 - `20:00` 全球海缆分布深度受制于海沟、洋流及火山等海洋地质，并高度重合于苏伊士运河等传统贸易咽喉，埃及等地由此成为全球信息咽喉。 - `22:30` 互联网路由遵循全局拓扑而非直线距离，霍尔木兹海峡或红海等咽喉海缆受损会迫使全球数据重新绕行，造成远端主干网拥堵与延迟外溢。 - `25:00` 现代海缆修复过程充满技术错位感：深海作业仍依赖 19 世纪式的铁锚抓钩海底盲捞，但拉上甲板后则需在无尘实验室内实施微米级光纤熔接。 - `26:30` 在 Meta 主导的非洲环线“2Africa”项目中，缺乏议价筹码的发展中国家为了获得宽带接入，被迫在本土数据主权与监管上作出妥协。 - `30:00` 中国企业华海通信凭借成本与工程优势深耕亚非拉海缆市场，但在欧美制裁与地缘阻断下，中美跨国海缆项目出现停摆与搁浅。 - `35:00` “灰色地带战争”威胁与非对称破坏风险急剧攀升，英国已派遣常态化军舰巡逻护缆，台湾沿岸 15 条海缆也被视为潜在冲突中的脆弱节点。 - `36:30` 现代海缆登陆点高度承袭 19 世纪电报站的旧址导致单点暴露，迫使海湾国家重新投资穿越叙利亚与伊拉克的陆地数据走廊以绕开海峡。 - `40:00` 生成式 AI 与超大规模数据中心的爆发彻底打破了前几年关于“带宽过剩”的悲观论调，驱动新一轮全球跨洋海缆资本开支繁荣。 - `42:47` 跨国顶级投行未曾自建跨洋专属专线，全球唯有中美等军事强权才会铺设与公共互联网完全物理隔离的深海专网。 - `45:00` 海缆建设周期与电力能源基础设施高度相似，早期的“产能过剩”往往只是技术概念超前与资本节奏之间的微小时间错位。 ## Detailed walkthrough ### `00:00` 互联网背后的物理骨架：被忽视的深海线缆网络 在播客开场，主持人 Tracy Alloway 和 Joe Weisenthal 探讨了现代数字化世界背后的物理现实。Tracy 让 Joe 在笔记本上打开一张大英百科全书（Encyclopaedia Britannica）的海底光缆分布图。虽然 Joe 一开始打趣自己弄丢了链接，但他立刻认出那些纵横交错的线路极度形似全球海上航运航线（Shipping lanes）。尽管海底光缆并不严格贴着主要航道敷设，但由于它们连接的都是各大全球经济中心和沿海大都市，整体布局呈现出高度的重合感。 ```mermaid flowchart LR accTitle: 数字与物理基础设施的对应关系 accDescr: 展示从用户终端的无线错觉，穿透到数据中心，再到深海物理光缆的层级传递流程。 subgraph User["用户终端体验"] A["智能手机 / 笔记本电脑"] B["看似'轻盈'的无线与 AI 概念"] A --- B end subgraph Backbone["物理支撑基础设施"] C["重资产、高耗电的数据中心"] D["跨洋海底光纤电缆"] C --- D end User -->|"依赖底层物理传输"| Backbone ``` 这种实体感与人们对 AI 或互联网的日常认知形成了强烈的反差。人们往往倾向于将网络视为某种漂浮在空中的轻盈数字存在，但现实中它完全依赖于庞大的重资产物理基础设施——无论是消耗海量电力和大宗商品的数据中心，还是蜿蜒在深海海床上数千英里的光纤电缆。英镑兑美元汇率在金融界被称为“Cable”，正是源于一百多年前横跨伦敦与纽约之间铺设的大西洋海底电报电缆。至今仍能跨越大洋铺设一根连续不断的线路，依然令人深感震撼。 - **物理装甲与脆弱性**：部分靠近海岸或浅海区域的海底线缆配有厚重装甲以防外力破坏，但随着近年地缘政治动荡，海底光缆屡屡被视为国家级冲突中的潜在脆弱点。 - **海底剪线假想**：Joe 半开玩笑地提到，如果有人想切断某地的互联网，理论上只需把连向那里的所有电缆切断，甚至戏称可能出现派带有机械爪的“无人机龙虾”（drone lobsters）潜入深海剪缆的场景。 --- ### `05:00` 写作契机与汤加火山喷发的现实启示 本期嘉宾 Samanth Subramanian 是新文化与政治杂志《Equator》的执行主编，也是新书《海浪之下的网络：连接我们世界的脆弱缆线》（*The Web Beneath the Waves: The Fragile Cables That Connect Our World*）的作者。谈及为何研究这个横跨科技与地缘的冷门领域，Samanth 回忆起科幻作家尼尔·斯蒂芬森（Neal Stephenson）在 1990 年代为《连线》（Wired）杂志撰写的一篇长达四万字、几乎占据整期杂志的里程碑式长文《大地母亲主板》（*Mother Earth Motherboard*）。 斯蒂芬森在文中自封为“黑客游客”（hacker tourist），周游全球探访深海电缆的登陆点、铺设现场和维修队伍，描绘了当时方兴未艾的互联网蓝图。2012 年，Samanth 在去杂货店买菜的路上用手机阅读这篇长文，突然被一种强烈的荒诞感击中：自己之所以能在手机屏幕上顺畅加载这篇文章，完全仰赖脚下深海中某根冰冷的光缆正在实时传输数据。 触发他正式下决心成书的事件发生在 2022 年，南太平洋岛国汤加（Tonga）附近爆发了大规模海底火山喷发。剧烈的海底泥石流和滑坡直接切断了汤加唯一一条通往外部世界的国际海底光缆，令该国瞬间陷入“数字黑暗”。这促使 Samanth 展开深入调查：当一个现代社会失去网络连接会发生什么？现今铺设海底光缆的参与方、出资方和各国政府的角色发生了怎样的转变？ - **《大地母亲主板》的余韵**：在 1990 年代斯蒂芬森写下报道时，互联网尚未经历后续的大数据爆炸，但已经揭示了它对类似 19 世纪电报线缆物理架构的依赖。 - **汤加案例的样板意义**：汤加断网成为了绝佳的极端压力测试案例，直接促成了全书的调研与撰写。 --- ### `07:20` 从铜线电报到“光分复用”的技术跃迁 针对已故参议员特德·史蒂文斯（Ted Stevens）曾因将互联网形容为“一系列管道”（a series of tubes）而广受嘲讽的公案，主持人与嘉宾都认为公众对他其实有所误解。当普通人拿着无线手机时，很难感知到实体连接，但实际上互联网的骨干就是实打实的“线缆管道”。不过，现代深海光缆与当年连接纽约和伦敦的跨大西洋电报缆之间，存在着翻天覆地的技术代差。 早期的电报电缆核心是厚重的铜导线，依赖电流脉冲传输信号，到达对端后再进行解码。而现代海底光纤则是高纯度玻璃制成的工程奇迹，位于深海最深处的光纤纤芯甚至只有人类头发丝那么细。信息通过激光脉冲射入玻璃核心，在管壁之间发生全反射高速向前弹跳。为了在单根细丝中传输海量数据，行业普遍采用了波分复用（Wave Division Multiplexing, WDM）技术：将多路数据调制到不同频率的光波上并行发送，光线穿过大洋到达终点后，接收器再将各个频段的光波分离并解码复原为数据流。 ### `10:00` 海底光缆的铺设工艺与单一软件垄断 现代光纤技术呈指数级提升了单根海缆的数据承载能力，但这条产业供应链却极度集中。在全球范围内，有能力制造这种高精密光纤海缆的企业屈指可数，形成了一个极具技术壁垒的产能瓶颈。而在出资建造一条海缆时，流程极其严谨：资方首先会出资派遣专业的勘测船（survey ship）出海，对预定路线进行详尽的海底地形勘测——例如从伦敦途径里斯本、开普敦，一路延伸到阿联酋。勘测船返回并制定出最佳铺设路径后，真正的铺缆船（cable-laying ship）才会启航，在漫长的周期内沿既定坐标极其缓慢地将线缆放入深海。 铺设速度的控制是一门极其脆弱的平衡艺术：如果船速过快，悬垂入海的海缆承受过大拉力可能会直接断裂；如果船速过慢，线缆在海底堆积又会造成严重的松弛（slack）与材料浪费。令人震惊的是，控制这套精密运动的核心技术掌握在极少数人手里——全球几乎所有铺缆船使用的导航、航线计算及航速控制软件，竟然全部来自同一家独家软件供应商，它直接决定了线缆以怎样分毫不差的姿态沉入海底。 针对铺缆船的基本构造，主播好奇地发问：无论维多利亚时代铺设最初的电报电缆，还是如今的高科技时代，其本质是不是依然只是一艘装载着巨型卷轴（spool/drum）并随着航行慢慢放线的船？Samanth Subramanian 证实了这一点：核心机制确实从未改变，依然是将海缆缠绕在巨型卷筒上装船出海。唯一的重大差异在于，现代光纤海缆远比当年裹着厚重金属和绝缘胶的电报缆更轻便，因此同一艘工程船如今能够搭载并布设里程长得多的线缆。 - **产能瓶颈**：全球仅有极少数制造商具备生产大容量海底光纤线缆的核心能力。 - **软件垄断**：全球用于计算铺缆船航线、航速及海缆张力的专业航行控制软件，几乎被单一公司完全垄断。 - **机械传统**：巨型卷筒（drum/spool）缓慢放线的基本作业模式自19世纪电报时代延续至今，但线缆自重的大幅减轻使得单船装载里程显著提升。 ### `11:47` 为什么卫星无法取代“管道”：海量数据的物理重力 已故参议员特德·史蒂文斯（Ted Stevens）曾因把互联网形容为“一系列管道”（a series of tubes）而广受嘲弄，但这一比喻在物理本质上其实非常准确。面对“既然无线与移动网络如此普及，为何不能彻底摆脱物理线缆全部通过蜂窝或卫星通信”的疑问，Subramanian 解释道，卫星通信虽然早已成熟——例如海事卫星电话以及 Garmin 等可下载天气数据的卫星终端——但它们根本无法应付人类日常吞吐的数据洪流。 将全球数据转移至近地轨道的想法是不切实际的，即使把地球近地轨道全部塞满卫星，其带宽容量也远远承载不下全人类的需求。现代互联网背后沉睡着庞大到令人窒息的数据实体：数以亿计用户的 Netflix 高清流媒体视频、全天候运转的 Zoom 视频会议、海量短信、跨国电话、毫秒级高频日内交易，以及全部托管在云端服务器上的海量办公演示文稿与企业数据中心。这些所谓“漂浮在云端”的数据，归根结底必须依赖实体光纤那近乎无限的物理带宽在各大洲之间高速流转。 ```mermaid flowchart TD accTitle: 现代互联网数据传输载体对比 accDescr: 对比海底光纤网络与卫星通信网络在容量和应用场景上的根本差异 subgraph UserData["全球日常数据需求 (海量级)"] D1["高清视频流 (Netflix)"] D2["企业会议与协作 (Zoom)"] D3["金融高频交易与云计算"] end subgraph Sat["卫星轨道通信链路"] S1["低/高轨通信卫星星座"] S2["局限: 带宽与吞吐容量存在物理瓶颈"] end subgraph Ocean["海底光纤主干网络 (Series of Tubes)"] O1["跨洋大容量光纤线缆"] O2["优势: 支撑全球绝大多数跨洲网络流量"] end UserData -->|吞吐量远超轨道承载上限| Sat UserData ==>|承担95%以上全球数据主干| Ocean ``` ### `13:16` 从国家电信联盟到科技巨头垄断：海缆资方四十年变迁 海底光缆由谁决定走向、由谁出资建造的权力格局，在过去数十年里经历了三次深刻的范式转移： 1. **1980至1990年代的国有电信财团时代（Consortium Model）**：当时全球多数电信运营商均为国有企业。若要铺设一条从伦敦经葡萄牙、穿过三个非洲国家并延伸至新加坡的长途海缆，沿线各国的国有电信运营商会组建联合体（consortium）。各方按出资比例认领建设成本，换取对应配额的带宽，再倒手销售给国内的消费者与企业。 2. **1990年代末至2000年代的私人投资驱动时代（Investor-led Model）**：伴随全球电信私有化浪潮，原有国有巨头被推向资本市场，投资人主导模式随之兴起。精明的私人资本会主动发现商业路线空白，在公开市场上募集资金独立铺设海缆，然后再向各大电信运营商分销出租带宽配额。 3. **近七八年以来的“Big Tech”自建时代**：以 Google、Meta、Amazon 和 Microsoft 为代表的科技巨头积累了无与伦比的现金流。一条从伦敦连接至纽约的大西洋海底光缆耗资大约在 5 亿美元上下——这笔费用对普通大众乃至小型主权国家而言是天文数字，但对市值数万亿美元的科技巨头来说仅仅是九牛一毛。 鉴于数据是科技公司的生命线，自建底层基础设施成了最高优先级的商业策略。Subramanian 在调研该书时发现，全球每铺设三条新海缆，就有两条是由这四大科技公司部分或全资拥有。这种资本集中不仅改变了互联网的物理版图，更在根本层面引发了对于数据隐私、国家数据安全，以及“究竟由谁来掌控人类接入互联网的权利”等一系列严重的地缘政治和数字主权担忧。 - **典型造价**：单条横跨大西洋（伦敦至纽约）的现代化海底光缆工程成本约为 5 亿美元。 - **掌控力剧增**：四大科技巨头（Google、Meta、Amazon、Microsoft）参投或控股了全球约三分之二的新建海底光缆项目。 ### `17:42` 铁路天然垄断与海缆冗余博弈：巨头间的竞合网络 在互联网基建初期，舆论常用19世纪的“铁路热潮”（railway booms）来类比其经济学特征。铁路往往具有极强的“天然垄断”（natural monopoly）属性：在纽约与芝加哥之间，若有五条互不相连、由不同公司运营的平行轨道紧贴在一起，从资本和土地利用上看是极度浪费且不经济的；普通乘客哪怕渴望多家铁路公司竞争来平抑票价，现实中也不可能在康涅狄格州的某座小镇与纽约之间铺上四条轨道。 然而，海底光缆的底层经济学与铁路截然不同，核心在于其对“系统冗余度”（redundancy）的极端依赖。一条横跨大洋的光缆哪怕被 Google 独资掌控，且假设其仅用来传输自家的 Gmail、Google Meet 等内部流量，Google 依然有极其强烈的经济与安全动机去向竞争对手购买备用容量。 因为海缆脆弱易损，一旦遭遇地震、滑坡或被船锚刮断，业务就将陷入瘫痪。因此，现实中不仅同一路线上并存着多条由不同实体铺设的平行物理光纤，巨头之间更存在着错综复杂的互保机制——Meta 会购买 Google 海缆上的备用带宽作为灾备，Google 也会反向采购 Meta 旗下线缆的容量。这种对零停机时间的极端追求，打破了单一管线的天然垄断，在深海之下编织出了一张高冗余、多重互锁的竞合网络。 ### `20:00` 全球海缆冗余网络与严苛的海洋地理现实 面对可能发生的中断事故，互联网底层架构其实建立在一套高度依赖冗余的防御机制之上。全球大洋底部铺设了大约 500 到 550 条海底光缆，它们绝大多数密集分布在经济活动最活跃的航道与区域：从西欧通往美国东海岸、连接中国与东南亚的航线、波斯湾乃至印度半岛。这些航道之所以光缆密布，根本原因就是为了容灾备灾。据统计，全球每年大约会发生 100 起海缆切断事故，其中绝大多数并非敌意破坏，而是纯粹的意外——比如路过货轮抛锚时刮断线缆，或是深海拖网渔船在海底作业时不慎挂住并扯断光缆。然而，普通网民几乎从未感知到一年高达百次的海缆断裂，正是因为冗余系统在发挥作用：一方面有穿透大陆的陆地光纤（如美加之间的陆上光缆分担流量），另一方面跨大西洋等繁忙水域本身就并存着足够密集的平行光缆，一旦某条受损，相邻线路立刻就能承接被切断的数据负荷。 海缆的走向与分布在很大程度上完全受制于地球的物理与地缘特征，其路径与传统海上航道高度重合。首先是极端复杂的深海地貌：大洋底部并非坦途，而是布满了深渊、海沟、湍急的洋流以及陡峭的海底悬崖，铺设作业必须小心翼翼地绕开这些险境。例如汤加恰好坐落在太平洋“环太平洋火山带”上，其附近的海底火山活动曾多次摧毁当地海缆，使得这个岛国在自然灾害面前显得格外孤立脆弱。 除了深海地形，水面上的地理瓶颈同样决定了光缆的命运。从亚洲前往欧洲，无论是货轮还是数据光缆，经由苏伊士运河穿行都远比绕道好望角经济得多。正如作家尼尔·斯蒂芬森（Neil Stephenson）在知名散文中所描绘的，当他走访埃及亚历山大港时，看到大批光缆在此交汇穿梭，连接欧亚大陆，此时的亚历山大港俨然成了全球信息的超级中枢与集散库，与古代著名的“亚历山大图书馆”形成了奇妙的历史呼应。埃及由此成为全球网络不可或缺的咽喉枢纽。 - 全球在运行的海底光缆总数约为 500 至 550 条。 - 全球每年发生约 100 起海缆断裂事故，主要诱因为船只锚泊与深海拖网作业。 - 铺设路线高度受制于洋底海沟、悬崖、洋流以及海底火山活动。 ### `22:30` 地缘咽喉与数据路由：霍尔木兹海峡的连锁反应 地缘战略要道对海缆安全的考验正变得愈发严峻。红海与霍尔木兹海峡等传统航运瓶颈，如今也是信息洋流的狭窄咽喉。如果蓄意破坏者企图瘫痪全球网络，只需派遣水下无人设备潜入这些水域，就能将横截在海床上的关键光纤逐一切断。当主持人问及，若仅切断霍尔木兹海峡中连接沙特与伊朗等国的局部光缆，影响是仅限于波斯湾沿岸还是会波及全球网络时，萨曼斯·苏布拉马尼安解释了互联网深层路由机制的非直观性：网络世界并不像传统陆地高速公路那样“两点之间直线最近就是最优解”。 在真实网络环境中，路由器在毫秒级内所做出的路径决策取决于瞬时流量与网络拓扑。举例而言，一位身在伦敦的用户向葡萄牙的服务器发送请求，数据包的最佳路径完全可能不是走海缆直达，而是先经由陆地穿过法国，再转运至葡萄牙；同样，一位沙特阿拉伯的 Gmail 用户，其个人邮件数据往往并不储存在本地，而是存放在西欧甚至美国的超大规模数据中心里。这意味着，哪怕只是特定地缘区域的一组海缆被切断，其产生的外溢效应也会迫使跨国数据流量在全球剩余的信道中不断寻找备用路径进行重新路由，进而加剧其他主干网的拥堵和延迟，牵一发而动全身。 ```mermaid flowchart TD accTitle: 复杂网络路由决策与断缆连锁反应 accDescr: 展示当区域海缆受损时，数据包如何基于全局服务器分布与网络拓扑被迫在冗余路径中重新绕行。 subgraph UserNodes["终端用户与数据请求"] U1["沙特用户 / 本地请求"] U2["欧洲用户 / 跨国请求"] end subgraph ChokePoints["地缘咽喉风险区（红海/霍尔木兹海峡）"] C1["区域核心海缆"] D1["潜在风险：锚损 / 蓄意破坏剪线"] C1 -.-> D1 end subgraph GlobalServers["远端数据中心与全球节点"] S1["西欧 / 美国数据中心 (如 Gmail 存储)"] S2["备用陆缆与跨大西洋海缆"] end U1 -->|"请求数据"| C1 C1 --"若遭遇切断"--> S2 U2 -->|"绕道陆缆（如经法国）"| S2 S2 --> S1 ``` ### `25:00` 维多利亚时代的打捞术与现代光纤接续：硬核的海缆维修 谈及海底光缆的抢修方式，现实工艺展现出了一种强烈的时代错位感。公众往往想当然地以为现代海缆维修会大量出动高精尖的无人潜水器（UUV），但苏布拉马尼安指出，其实际作业流程在很大程度上依然保留着 19 世纪维多利亚时代修补电报线缆的“老派”作风。一旦定位了故障区间，抢修船便会驶向茫茫大海中推测的断裂坐标。抵达后，船员会向海里抛下一个沉重的铁锚抓钩（grapnel hook），让它沉入冰冷的海床并贴着泥沙一路拖曳，试图盲钩住细弱的光缆。这种打捞常常遭遇挫折：好不容易感觉到抓钩咬合受力并将其拉上海面，拉上来的却可能只是一堆海底垃圾或废弃物，船员只能将其扔回海里重新抛钩拖网。 然而，一旦那根纤细的光缆被真正拉上甲板，现场就会瞬间切换回极致尖端的现代科技场景： - **船载精密实验室**：修缆船内建有高规格的无尘室（clean room），配备极为精密的抗摇摆自适应稳定平台，即使外部海况波涛汹涌，实验室内也能以惊人的稳定性运作。 - **高难度光纤熔接**：工程师在微观尺度下将断裂的玻璃光纤一根根重新剥离、精准对准并进行微米级熔接。 - **反复测试与复位**：完成熔接封装后，工程团队必须对光信号传输质量进行连环严苛测试，直到确保万无一失，才小心翼翼地把海缆重新沉放回海床原有的经纬度坐标上。 ### `26:30` 科技巨头的海缆霸权、地缘博弈与 2Africa 的妥协 在技术脆弱性之外，海缆产业更深层的危机源自其资本与地缘垄断。当前海缆铺设领域最大的软肋之一，莫过于四家美国科技巨头对全球基础网络投资的绝对主导地位。互联网的物理实体正日益屈从于这些巨头的商业意志——巨头想把网络修到哪里，哪里才能接入高速世界，这在很大程度上削弱了互联网原本去中心化、开放平等的构想。 以全长绕非洲大陆一圈、堪称全球最长海缆的“2Africa”项目为例，该工程几乎完全由 Meta（Facebook 母公司）出资并主导。这赋予了单一商业实体决定西非哪些国家能够获得登陆站点的绝对权力。对于急需互联网基础设施来拉动经济发展、在数字化进程中处于严重匮乏状态的西非国家（如科特迪瓦）而言，这构成了一个残酷的地缘政治困境： - 面对 Meta 递来的橄榄枝，发展中国家政府理论上也希望效仿西欧国家，开出强硬的数据主权保护条件，要求本国公民的数据必须存储在境内的本土服务器上； - 但由于缺乏足够的议价筹码与替代方案，如果提出严苛的监管要求，Meta 完全可以轻描淡写地选择放弃在该国靠岸，直接绕道邻国； - 这种悬殊的权力不对称，逼迫处于网络边缘的国家不得不牺牲公民的数据隐私与安全底线，以换取国民急需的宽带接入，“开放自由互联网”的乌托邦幻象在此被彻底击碎。 即便是在偏远如太平洋岛国基里巴斯（Kiribati），海缆也成为了大国地缘博弈的角力场，美、澳、日三国曾联合出资为其筹建原本定于 2025 年投入运营的海缆。而在亚非拉发展中国家，面对欧美巨头的强势垄断，中国力量也早已深度介入：华海通信（HMN Tech，前身为华为海洋）作为从华为剥离出来的深海工程企业，已经在海缆铺设领域深耕十余年，与美国的 SubCom 等老牌海缆铺设巨头同台竞技，凭借更具竞争力的成本优势和工程能力，为全球发展中地区提供了传统西方科技巨头之外的关键替代选项。 ### `30:00` 中美地缘博弈与互联网“双轨分叉”的隐忧 在地缘政治格局深刻重塑的背景下，海底电缆行业早已无法偏安于商业逻辑之内。中国公司如华海通信（HMN Tech，前身为华为海洋）不仅能够制造电缆，还能为企业客户出海铺设，且报价通常比欧美竞争对手更具成本优势。然而，自从首届特朗普政府将中国科技企业列入制裁清单以来，全球企业如果选择与华海通信或华为合作铺设海缆，就必须承受触犯制裁令及招致多重惩罚措施的高昂风险。回溯至八九年前，行业内曾有一段堪称“田园诗般”的互信时期：西方公司与中国企业会结成跨国联合体（Consortium）共同分摊高额建造成本，合理规划谁来执行铺设计划并按需分配带宽资源。如今，这种合作模式已荡然无存，数个重大跨国海缆项目被迫搁浅或冻结——其中一条电缆甚至已经完全铺设完毕并沉入海底，却因美国政府严防中国企业参与而始终无法通电启用；另外两个项目在完成路由勘测与资金承诺后，也不得不全盘撤回。 这种博弈直接导致中美之间多年未曾铺设任何一条直连的海底光缆。尽管两国之间的数据流量已攀升至历史峰值，如今所有通信数据却必须绕道菲律宾或新加坡进行中转。从中国东海岸辗转至美国西海岸，这是一种极度漫长、低效且非理想的“大兜圈”。更令业内人士夜不能寐的，是全球互联网正在不可逆转地面临某种“分叉”（Bifurcation）的危险。这种现象并非基于安全容灾的“冗余备份”（Redundancy），而是一种纯粹因地缘对抗造成的“系统多余”（Superfluity）——当某个地区急需数据互联，而美方排斥中方参与、中方又另起炉灶时，当地就会出现两套平行的海缆系统。这直接撕裂了互联网最初构想中那种“自由、开放、无缝连接全世界”的理想承诺。 ```mermaid flowchart TD accTitle: 中美海底电缆地缘政治阻断与路由绕行 accDescr: 展示中美直连海缆受阻后数据被迫绕行东南亚及造成系统冗余的现状 subgraph Direct["传统预期路径（受阻）"] CN1["中国东海岸"] -- "直连海缆项目冻结/未获审批" --> US1["美国西海岸"] end subgraph Reroute["当前现实绕行路径"] CN2["中国东海岸"] --> SEA["中继节点（菲律宾 / 新加坡）"] SEA --> US2["美国西海岸"] end subgraph Bifurcation["地缘博弈导致的体系分叉"] Region["重叠铺设区域"] --> US_Cable["美方主导海缆系统"] Region --> CN_Cable["中方主导平行系统"] end ``` - **联合体模式受挫**：八九年前西方巨头与中国企业共同出资、共分带宽的商业联合体几乎停摆，多条已测绘或已铺设的光缆因制裁沦为无法点亮的“海底死缆”。 - **绕行成本高昂**：中美双边直连通道断连多年，所有海量数据流被迫南下绕经东南亚枢纽（菲律宾、新加坡），网络延迟与传输效率大打折扣。 - **功能冗余与分裂**：平行铺设的两套系统并非用于容灾备份，而是互不信任导致的重复建设，威胁着全球单一互联网的底层互通性。 --- ### `35:00` “灰色地带战争”与海底基础设施蓄意破坏的威胁 虽然学术统计与历史数据显示，绝大多数海底光缆受损事故源于渔船拖网或商船抛锚等意外，但如今业内与各国政府愈发清晰地感知到：恶意蓄意破坏的风险正在显著激增。面对这种潜在威胁，全球各国的防卫戒备等级被层层拉高。以嘉宾居住的英国为例，政府几年前便正式宣布，将常态化派出两艘海军军舰在不列颠群岛周边海域全天候巡逻，其核心保护目标不仅包含海底数据光缆，还涵盖跨国输电电缆及海底油气管道。 在欧洲波罗的海与斯堪的纳维亚半岛周边，各国海岸警卫队的戒备状态提升至冷战以来的最高点，防范俄罗斯舰船在深海开展所谓的“灰色地带战争”（Gray Zone Warfare）。在亚洲，台湾当局同样高度警惕，担忧一旦发生冲突，通往台湾沿岸的全部15条国际海底通信光缆会被作为第一击的目标彻底切断。这种担忧并非空穴来风，各国安全部门敏锐地意识到，对于潜在的恶意行为者而言，切断几根海底光缆不仅成本极低、隐蔽性极高，而且足以在数小时内让一个现代经济体的通信与金融结算陷入彻底瘫痪。 - **英国常态化护缆**：部署两艘专属军舰，重点监控本国周边海域的数据光缆、电力线缆与油气管线。 - **波罗的海警戒**：北欧与波罗的海沿岸国家针对“灰色地带作战”加强巡逻，严防非对称破袭。 - **孤岛断网危机**：台湾高度依赖接入本土的15条海底光缆，成为地缘战略推演中最脆弱的关键基础设施节点之一。 --- ### `36:30` 历史路径依赖与绕开咽喉要道的新数据走廊 面对脆弱的深海通道，最直观的防御方式似乎是构建冗余度，但这绝非多扔几根电缆那么简单。如果对比19世纪末期全球海底电报电缆走线图与现代海底光纤网络分布图，会惊讶地发现两者的重合度极高——除了近几十年在东亚和东南亚蓬勃兴起的新密集网络外，全球海缆的登陆点与百年前几乎如出一辙。这种强大的历史惯性背后，是商业成本与规模效应的驱使：在印度，几乎所有国际海缆都集中登陆在西海岸的孟买或东海岸的钦奈，因为一个世纪前的电报线路就在这里上岸，相应的陆上机房、数据中心及配套基础设施早已在此扎根，继续在老地点扩建显然阻力最小、成本最低。然而，这种过度集中也意味着致命的脆弱性。 真正的破局之道在于寻找新的登陆点并开辟全新的陆上或海上走廊。主持人提及，沙特阿拉伯、卡塔尔与阿联酋目前正斥巨资资助穿越叙利亚、伊拉克及东非的竞争性数据走廊，旨在从陆路绕过动荡逼仄的霍尔木兹海峡。长期以来，海缆之所以宁可冒险挤进霍尔木兹海峡也不走中东内陆，是因为叙利亚与伊拉克在过去二十至三十年间始终深陷战乱漩涡，在战火连天之地根本无法确保光缆的实体安全。但如今地缘现实的变化正在倒逼资本与政府重新审视并激活这些曾经不可行的替代路线，从长远来看，这种被迫的地理多元化正在重塑全球数据版图。 - **百年登陆点未变**：印度海缆高度扎堆于孟买与钦奈，承袭了19世纪大英帝国电报网的旧址，暴露出严重的路径依赖与单点脆弱性。 - **绕道霍尔木兹海峡**：中东海湾国家联合投资陆地过境网络，试图通过叙利亚和伊拉克打通连接欧亚的替代通道。 - **安全与战乱权衡**：过去几十年陆地战乱迫使运营商依赖狭窄海峡，而今深海风险升级反向推动了陆上数据走廊的重建。 --- ### `38:15` 卫星能否取代光缆？“无缆化”未来的神话与现实 随着5G技术的普及和低轨卫星互联网（如埃隆·马斯克的星链）的飞速迭代，外界频频产生一种技术幻想：未来的互联网是否能彻底抛弃笨重易碎的深海光缆，实现完全的“空中无线化”？针对这一畅想，萨曼斯·苏布拉马尼扬给出了斩截的否定回答：“我不认为业内有任何人在严肃探讨‘后光缆时代’。”尽管卫星技术在吞吐量与连接优化上取得了长足进步，但在处理跨大陆的工业级数据体量时，其物理带宽上限与光纤相比依然存在不可逾越的数量级鸿沟。 卫星通信的真正价值在于极端环境下的应急兜底与偏远补盲，而非替代骨干网。以汤加火山爆发为例，在海底光缆被火山喷发切断后的漫长16个月中，汤加的一个偏远外岛正是凭借卫星链路勉强维系着最基本的信息对外连线。但这仅仅是灾难之下的权宜之计，支撑现代全球文明日常运转的庞大金融数据流、流媒体与云端算力交互，依然绝对依赖于那些深埋在各大洋海底、纤细而脆弱的玻璃纤维。 ### `40:00` 卫星无法替代光缆：AI 狂潮与海底基础设施的加速扩张 在谈及星链（Starlink）等低轨卫星通信系统时，萨曼斯·苏布拉马尼安（Samanth Subramanian）指出，尽管卫星在特定偏远地区或紧急场景下表现抢眼，但它绝不可能彻底取代海底光缆。核心矛盾在于，人类对数据的吞吐“胃口”正以远超卫星扩容速度的指数级爆发。当前的 AI 浪潮尤其印证了这一点：海量的模型训练数据存放在集中式的服务器集群中，全球用户在笔记本电脑上随时调用 Claude 或 ChatGPT 等引擎。这种数据流动机制与交互形态在五年前根本不存在，给整个光缆行业带来了前所未有的新增数据负荷。因此，业界普遍清醒地认识到，用卫星替代全球数据基础设施纯属无稽之谈，海底光缆在相当长的时期内仍将是绝对支柱。 主持人询问，苏布拉马尼安的新书《浪潮之下的网络》（*The Web Beneath the Waves*）于去年十月出版，生成式 AI、超大规模云服务商（hyperscalers）和数据中心建设的激增，究竟如何改变了海底光缆的发展轨迹？苏布拉马尼安认为，AI 并没有从根本上改变发展方向，而是极大地“按下了加速键”。早在 AI 热潮之前，大型科技巨头就已开始自主投资铺设跨洋光缆；如今，这批巨头全数投身超大规模 AI 研发，随着训练数据中心的成倍扩建，跨区数据交换量呈爆发之势，铺设更多、更快速光缆的需求只增不减。 这种高涨的需求彻底扭转了行业数年前的悲观论调。大约五六年前，业内一度充斥着“光缆产能过剩”的忧虑，许多人认为既有管道已足够应付未来，短时间内无需再铺设新线缆，但 AI 繁荣将这一过剩假象彻底打破。苏布拉马尼安同时强调，海缆铺设高度依附于全球宏观经济周期。光缆铺设是极度资本密集型的项目，一旦遭遇经济衰退，融资收紧就会导致新项目断崖式下跌。例如 2000 年代初互联网泡沫破裂后，海缆行业陷入了长达近五年的停滞期，当时虽然预订了大量额外光缆，却因缺乏资金而无人敢动工敷设。海缆业的周期性波动始终存在，但在本轮周期中，只要 AI 革命的浪潮持续，为了支撑整套体系的运转，大规模铺设新缆的步伐就不会停歇。 ```mermaid flowchart TD accTitle: AI浪潮与宏观周期对海底光缆投资的驱动机制 accDescr: 描述AI技术爆发与宏观经济周期如何共同影响海底光缆的资本开支与铺设进程。 subgraph TechDriver ["技术与需求驱动"] AI["AI 繁荣与大模型调用<br/>（如 ChatGPT、Claude）"] DC["超大规模训练数据中心<br/>跨区域数据流激增"] Hyperscaler["大型科技巨头（Hyperscalers）<br/>自主加码铺设专属海缆"] AI --> DC --> Hyperscaler end subgraph MacroCycle ["宏观经济与历史周期"] Boom["繁荣期：巨额资本开支注入"] Bust["衰退期：融资枯竭与5年停滞<br/>（如 2000 年互联网泡沫破裂）"] Overcapacity["过剩错觉（5~6年前预判）<br/>被新一轮需求彻底打破"] end Hyperscaler --> Boom Boom -.-> Bust Bust -.-> Overcapacity Overcapacity --> AI ``` ### `42:47` 华尔街专网、大国军用暗网与“最萌”短光缆 主持人探讨了光缆的早期金融渊源：光缆最初正是为了满足金融交易的跨地域信息同步、消除市场定价时差而蓬勃兴起的；如今跨国金融机构通常会依赖更高安全性、更低延迟的专用私有网络，而非公共互联网。主持人由此好奇：诸如摩根大通（J.P. Morgan）这类在伦敦和纽约均设有大型枢纽的顶级投行，是否曾出资铺设专属且与公共网络完全物理隔离的跨洋海底光缆？ 苏布拉马尼安回应称，金融界尚未出现独资铺设专属跨洋海缆的先例。不过在陆上光缆领域确实发生过类似事件——迈克尔·刘易斯（Michael Lewis）在其畅销书《闪击者》（*Flash Boys*）开篇所描写的真实场景，正是高频交易机构为了争夺数毫秒的时间优势，自筹资金硬生生在芝加哥与新泽西之间凿穿山林、铺设直线光纤。 而在真正的大洋深处，唯一会投入巨资自行敷设专线、并对外部公共互联网实现百分之百物理隔离的实体，是世界上的各大军事强权： - **军用专属海缆系统**：美国军方与中国军方均拥有专属于本国部队的海底光缆网络，由专门的防务承包商负责铺设与维护，公共互联网流量绝对禁止混入其上。 - **最喜爱的趣味海缆**：当被问及个人最偏爱哪一条海底光缆时，苏布拉马尼安笑着分享了一条“可爱到甚至本不该存在”的短距离光缆。他翻阅新书确认，这是一条连接曼岛（Isle of Man）与北爱尔兰之间的海缆，全长仅约 41 英里（约 66 公里）。这条缆线之所以让他着迷，是因为它在如此短的跨度内敷设，极其纯粹地展示了现代光纤技术惊人的成本效益比。 ### `45:00` 节目复盘：巨头垄断、玻璃光波的奇迹与“周期错位”反思 访谈结束后，主持人乔·威森索（Joe Weisenthal）与特雷西·阿洛威（Tracy Alloway）展开了深度复盘与延伸讨论： 特雷西感叹，全球海底光缆的发展格局恰恰是宏观经济演变的一面镜子。过去的海底工程由多国电信企业组成的联合体（Consortium）共同分摊成本与风险；而如今，整个生态已被极少数（通常是四家）超大规模科技巨头实质性接管与主导。此外，军方拥有完全平行的专属海底网络这一事实，既令人震撼又符合地缘政治逻辑。 乔则从纯工程与物理视角赞叹海底光缆是“人类最伟大的奇迹之一”——光束进入脆弱细微的玻璃纤维通道，在浩瀚冰冷的大洋深处千回百转地全反射弹射，跨越数千英里后，竟能分毫不差地在对岸解码还原为初始信息结构。这种物理实现令每一次日常跨国网页浏览都充满了超现实感。 针对海缆建设的周期性起伏，两位主持人提到了与现代电力能源行业的惊人相似性： - 当社会需求进入平稳平台期时，基础设施投资便陷入停顿； - 一旦 AI 等革命性技术突袭，电力负载与数据带宽需求骤然飙升，人们才惊觉基础设施扩张的紧迫性。 乔回顾了 2000 年前后的互联网泡沫时代。2002 年时，华尔街和舆论曾认定当时铺设的巨量光纤是“永远用不完的浪费”；然而仅过两三年，带宽就被迅速填满并重启建设。这证明当年科技先驱们对“互联网未来将变得无比庞大”的预判完全正确，彼时的所谓“过剩”，本质上只是一场“微小的时间错位（timing mismatch）”。特雷西对此精辟总结道：“一项技术完全可能在某一个特定时间点上，既处于概念被低估的状态，又在资本市场上被高估（Technology can be underhyped and overvalued at a single point in time）。” 节目尾声，特雷西与乔致谢了嘉宾萨曼斯·苏布拉马尼安，并公布了制作团队成员（制片人 Carmen Rodriguez、Dashiell Bennett 以及 Kel Brooks）与官方 Discord 讨论社群信息，随后播出了包含 4imprint 定制礼品、The Hartford 商业风险管理与保险、以及美属维尔京群岛（USVI）旅游局的相关赞助广告。 ## Notable quotes > `07:20` It's a 'series of tubes.' — Ted Stevens (quoted) > `00:00` drone lobsters — Joe Weisenthal > `05:00` hacker tourist — Neal Stephenson (quoted) > `05:00` Mother Earth Motherboard — Neal Stephenson (quoted) > `05:00` The Web Beneath the Waves: The Fragile Cables That Connect Our World — Samanth Subramanian > `11:47` a series of tubes — Ted Stevens (quoted) > `35:00` Gray Zone Warfare — Samanth Subramanian > `38:15` 我不认为业内有任何人在严肃探讨‘后光缆时代’。 — Samanth Subramanian > `45:00` timing mismatch — Joe Weisenthal > `45:00` Technology can be underhyped and overvalued at a single point in time — Tracy Alloway ## Mentioned Samanth Subramanian, Tracy Alloway, Joe Weisenthal, Ted Stevens, Neal Stephenson, Wired, Encyclopaedia Britannica, Equator, Google, Meta, Amazon, Microsoft, Netflix, Zoom, Garmin, HMN Tech (华海通信), SubCom, Starlink, Claude, ChatGPT, Michael Lewis, J.P. Morgan, 2Africa, Carmen Rodriguez, Dashiell Bennett, Kel Brooks, 4imprint, The Hartford ## Video description <p>In 2006, then-Senator Ted Stevens coined an infamous term for how to understand the internet: It's a "series of tubes." The funny thing is, that's a fairly accurate description. Underneath the world's oceans, miles and miles of fiber optic-cables send packets of information from one location to the next, serving as the backbone of the internet as know it. This infrastructure is delicate, too: Memorably, a 2022 volcanic eruption cut off the island of Tonga from web access for an extended period of time. Journalist Samanth Subramanian is the author of <em>The Web Beneath the Waves: The Fragile Cables That Connect Our World</em>, a book that explains, in detail, that the internet is not, and has never been, truly weightless or wireless. In fact, the system in place right now is pretty old school and resembles the telegraph cable network of yore. We talk to Subramanian about the strange contradictions of the undersea cable system, how much basic marine geography &mdash; like the Strait of Hormuz or the Suez Canal &mdash; informs where cables are laid, and how hard it is protect this vulnerable and vital infrastructure.</p> <p><a href="http://bloomberg.com/subscriptions/oddlots" data-link-cookie="1" data-destination="http://bloomberg.com/subscriptions/oddlots">Subscribe to the Odd Lots Newsletter</a><br>Join the conversation: <a href="https://discord.gg/oddlots" data-destination="https://discord.gg/oddlots">discord.gg/oddlots</a></p><p>See <a href="https://omnystudio.com/listener">omnystudio.com/listener</a> for privacy information.</p> [Source](https://podtrac.com/pts/redirect.mp3/tracking.swap.fm/track/UVBrz8bN8aM2Xe47PEPu/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/8a94442e-5a74-4fa2-8b8d-ae27003a8d6b/d05e23d4-18e1-442f-8953-b4480146b311/audio.mp3?in_playlist=982f5071-765c-403d-969d-ae27003a8d83)

[7:20] Ted Stevens (quoted): It's a 'series of tubes.' [0:00] Joe Weisenthal: drone lobsters [5:00] Neal Stephenson (quoted): hacker tourist [5:00] Neal Stephenson (quoted): Mother Earth Motherboard [5:00] Samanth Subramanian: The Web Beneath the Waves: The Fragile Cables That Connect Our World [11:47] Ted Stevens (quoted): a series of tubes [35:00] Samanth Subramanian: Gray Zone Warfare [38:15] Samanth Subramanian: 我不认为业内有任何人在严肃探讨‘后光缆时代’。 [45:00] Joe Weisenthal: timing mismatch [45:00] Tracy Alloway: Technology can be underhyped and overvalued at a single point in time

[0:00] Joe Weisenthal: It blows my mind that like over 100 years ago... they were able to string a continuous line from London to New York... I still find it hard to believe that they can actually have one length of cord that goes undersea this far. [25:00] Samanth Subramanian: It gives the lie to the old sort of utopian model of a free open Internet [39:30] Samanth Subramanian: However much satellites improve, our appetite, our sheer thirst for data is just going to improve exponentially... nobody thinks that we can replace all our data infrastructure with satellites. I think cables are here for a good long while yet.

[0:00] Whatever your goal, trade show giveaways, client gifts, or team gear, 4imprint has the promo products to match. With thousands of options, from apparel and drinkware to tech and totes, it's easy to find the right fit for your brand and budget, with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain. The thing about AI for business? It may not automatically fit the way your business works. At IBM, we've seen this firsthand. But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM. As industries evolve faster than ever, companies need an environment that accelerates strategic growth, and Michigan delivers on that promise. From emerging startups to global enterprises, Michigan offers what executives value most a resilient innovative ecosystem diverse communities that attract top talent and a quality of life that supports work balance With our unified Team Michigan approach businesses scale faster and compete at the highest level Michigan, pure opportunity. Seize your opportunity at michiganbusiness.org. Bloomberg Audio Studios, podcasts, radio, news. Hello and welcome to another episode of the All Thoughts Podcast. I'm Tracy Allaway. And I'm Jill Weisenthal. Joe, I am going to send you a link. Okay. You're going to open it on your laptop, which we have expressly for this purpose. And tell me what you see. Okay. Okay, I'm putting it... This is a fun experiment. We've never done anything like this before. I know, we've never been so prepared for a podcast as to have a link ready to go. It's a map of undersea... It's actually the Encyclopedia Britannica page of undersea cables. And there's a really old fashioned looking map here. It looks pretty cool. What does that remind you of? Oh, now I lost the line. I knew this wouldn't work. Shipping lanes Yes Oh okay good Yes that the right answer I got the right answer Yeah so I been looking at these maps and they look remarkably like shipping lanes even though upon further reading it turns out that subsea cables do not precisely follow actual major shipping lanes But since most of them are going to and from a major economic center, a city or something, they look a lot like shipping lanes. And the reason I bring this up is because I think with everything technological nowadays, whether it's AI or the basic Internet, there's a tendency to think of it as this very ephemeral digital presence in our lives. But of course, as we've been discussing on a couple of podcasts recently, there is an incredible physical architecture, which is the source of all these things, whether it's subsea fiber optic cables that kind of look like shipping lines or massive data centers that cost a lot of electricity and commodities to produce. Undersea cables is one of the most fascinating things to me because there's been a cable right between London and New York for over 100 years. I know. And that blows. That's actually why the UK pound sterling rate is called cable. That's why it's nicknamed that. It blows my mind that like over 100 years ago, whenever it was when they first had the telegraph that they were able to string a continuous line from London to New York That blows my mind that they could do that whenever that was like a century ago or whenever But it blows my mind that they can still do it today Like it actually, I still find it hard to believe that they can actually have one length of cord that goes undersea this far. Even with all the technologies, I still can't wrap my head around how they do that. Yeah, and some of those cable lines, I think are armored in various ways to protect them. They're talking about doing like more land fiber optic cables now as well. There's a lot that we should discuss, not just because this topic is incredibly interesting and we've been meaning to for a long time, but also because with all the geopolitical volatility that we're seeing nowadays, you always hear subsea cables coming up as a potential source of vulnerability. Totally. You hear about attacks on them from time to time. And I guess there's probably some redundancies and other ways to route around. But it is weird to think that, like, in theory, you could just like cut off some part of the world from the Internet by clipping all the cables. At least I get that. You could send out little drone lobsters and they like cut the cables on the deep sea bed with their little lobster claws. That's probably literally happening. OK, we should actually speak to someone who knows about this. We do, in fact, have the perfect guest. Again, someone we wanted to speak to for a very long time. We're going to be speaking with Samanth Subramanian. He's the author of... [5:00] The web beneath the waves, the fragile cables that connect our world. He's also the acting manager editor of Equator, which is a new magazine covering politics and culture. Sounds very cool. Saman, thank you so much for coming on All Thoughts. Thank you for having me. I should have said in the intro that you're actually the author of numerous books on kind of, I would say, disparate subjects. So I'm very curious why you decided to cover something like subsea cables. Well, I mean, there's a long answer and a short answer. And the long answer has to do with this essay that I read more than 10 years ago. This was an essay by the science fiction writer Neil Stephenson. And he had written this piece in Wired magazine sometime in the 90s, 40,000 words long. It took up like virtually the entire magazine called Mother Earth Motherboard. And that essay had Stephenson, you know, he sort of cast himself in the role of what he calls a hacker tourist. And he goes around the world, kind of looking at places where these subsea cables land and are installed into the earth or let into the sea or repaired. And he meets these like odd people who do this kind of work or did this kind of work back then. And he kind of traces through all of this a picture of the nascent Internet I mean the Internet was already there but I guess it wasn as big as it was now obviously But you know it wasn even as big as it was in 2005. There was a lot of big data explosions yet to come, but it was clearly something that was burgeoning. And I think Neil Stephenson did such a good job, even in the 1990s, of reminding people that the internet is reliant on this physical infrastructure, this actual cabling that looks remarkably similar to telegraph cables from the 1800s, although, of course, there are big differences. So he wrote this essay in the 1990s. I read it in 2012, I think, and I remember being so entranced by the story that I had to step out to get groceries and I kept reading on my phone. I read in the shop. I read on the way back. I was walking to and from there. And of course, it struck me that the only reason I was able to load it on my phone as I was walking was because the internet was carrying this story to me through some undersea cable or another. And then fast forward to a few years ago when there was a big volcanic eruption off the coast of Tonga which is an archipelago in the South Pacific And Tonga lost connection to its only international subsea data cable The mudslide and landslide that kind of came out of the volcanic eruption underwater severed this line and Tonga was sort of plunged into a kind of internet darkness. And that started to make me think, well, what is it like today for a country or a society or an economy to live without the internet for even a brief period? What are these cables like today? How have they changed from the time that Neil Stephenson wrote this essay in the 1990s? Who lays them now? Who owns them now? Has the funding changed? Do governments play a different role, similar role? And I kind of wanted to find all of this out. And I wanted to use Tonga as this kind of little test case of what it was like for a country to live without the Internet for a while. And that's why I pitched the book. so there's that famous phrase from ted stevens the senator that the internet is a series of tubes yeah it is though i never understood why everyone made fun of the guy that's literally what we're talking about it is a series of tubes okay maybe they're like pipes or cables but like it never really seemed that wrong i think the confusion comes with wireless right where you're holding your cell phone and it's like well my cell phone isn't actually attached to anything at the moment so it hard for people to wrap their heads around i always thought he was unfairly maligned Talk to us about the process the very simple explanation of how a long undersea cable is laid and how much is it fundamentally similar or different to when that first famous telegraph cable was laid between New York and London Well, the cables themselves are very different. I mean, the first telegraph cable was made of copper and you kind of send pulses of electricity through it and that comes out of the other end and it's kind of decoded. The modern fiber optic cable is a real technological marvel. I mean, I haven't stopped marveling at it since I started researching this book. The best cables are the ones at the bottom of the ocean are just a hair thick, just literally the thickness of a human hair. They're made of highly purified glass. And down that cable, you send sort of little pulses of light, lasers, essentially. And they kind of bounce around the inner walls of the glass and they come out of the other end and you can decode them. And what they do these days, there's a long name for it called wave division multiplexing, where they send different frequencies of light encoded with different streams of data. And so that kind of bounces around the glass at various speeds. And then it comes out of the other and you kind of suck it out and you code it all back together and you kind of read the information out that way. And that has kind of, [10:00] exponentially increase the amount of data that a cable can carry. You know, these cables are produced by just a handful of companies around the world. So there's a little bottleneck over there in terms of technological capacity to produce these cables. And then there's people who find the money for these cables. So that used to include state-owned telecom companies. Then it included private investors sort of raising money from a bunch of places. now, as we'll probably talk about, it's mostly big tech companies paying for it out of their own wallets because they can afford to. And then they would fund a survey ship to go out and see what the best route for the cable would be from, say, London to Lisbon to Cape Town to, I don't know, the UAE. And then the survey ship would come back and give you the best possible route. And then a cable laying ship would go out, make multiple trips and just slowly, very slowly lay the cable along exactly that route. And you have to do it that way because too fast and it might snap too slow and there's too much slack in the cable. And so you really need to kind of find this optimum speed at which to travel. And it's really funny. There one company in the entire world that makes the kind of software that all these cable ships use that determine the ship path the ship speed all of this stuff just so the cable is laid just so Just to be clear quick question Is it basically, whether we're talking about the Victorian era cable or now, is it a ship with a giant spool, basically, that slowly unwinds as it moves? That's basically, yeah. I mean, that hasn't changed. I mean, there's a couple of other things that haven't changed, but that definitely hasn't changed. They call them a drum or they call it a spool, but essentially wind the cable around the spool and then you load it on board and then you set off into the high seas. So that part of it hasn't changed. Of course, the cable is much lighter now, so you can load a lot more cable onto it because it's not sort of thick metal. Sorry, I realized we kind of skipped ahead, but for people who do like to malign Senator Stevens on his tubes comment, why is it that we can't just send data all through cellular means? Why Why do we need fiber optic cables at all? You can definitely. I mean, you can use satellites and in fact, satellite phones have been around for a long while. Garmin makes all of these satellite enabled devices and you can download weather data and other stuff onto it. It's just that the volume, the sheer volume of data that we ingest on a daily basis you couldn fill Earth orbit with enough satellites to take all that data up I mean think about what it means It means everybody Netflix streams and everybody Zoom calls and everybody's texts and phone calls and day trades and PowerPoint presentations that live on the cloud and data servers that serve other kinds of things. Everything is essentially on a cloud somewhere. And there's so much data out there that there's not enough satellites that good process. Okay, that makes sense. So the other thing I really want to ask you is who actually makes these decisions about what gets laid and where and who's financing it? I know you mentioned that big tech nowadays pays for most of it, but I imagine that must have changed throughout time, right? If we're thinking back to that first transatlantic cable, maybe it was wealthy industrialists trying to do something nice for the world. I don't know. Maybe it was governments pressing forward with it, more likely it was private companies. But the role of who's funding and planning these things must have changed throughout time. It definitely has. I mean, in the 80s and 90s, and I mean, I mean, the 1980s and 1990s, a lot of these telecom companies around the world were state owned And so they would kind of figure out that they needed a cable to run from let say London to Portugal to three African countries and then onwards to Singapore as an example So the telecom companies in each of these governments would come together and they would say well, look, this is how we want to lay the cable. This is how much money we can put up. In return, we get this much bandwidth and we can kind of sell that on to our customers. And so these companies would form essentially like a consortium. And that was the consortium model was popular for a very long time. And in the 1990s and the early 2000s, you saw this wave of privatization around the world. All of these telecom companies were hived off into private companies. And suddenly, at the same time, there was also this growth of the investor-led model. Like an investor would say, this cable looks like a good business idea. I will go out and raise money. And then I will come back to these telecom companies and say, look, I'm willing to lay this cable. How much bandwidth do you want to buy from me? And they would kind of apportion bandwidth in that way. And so that model prevailed for like much of the 2000s. And I think it's only about seven or eight years ago that big tech companies, by which I mean Google, Meta, Amazon, and Microsoft, A, grew so cash rich that they could afford these cables. So, for example, a transatlantic cable now from London to New York would cost about $500 million. [15:00] which is a lot of money still for you and me, but maybe not that much for Google. And so they started to fund these cables because rightly or wrongly, they thought, look, data is the lifeblood of our business and it makes more sense to build this infrastructure ourselves. And so now we got to the stage or when I was researching this book, it got to a stage where two out of every three new cables were being funded and owned either in part or in full by one of these four tech companies. and that has enormous sort of implications for data privacy, data security, I mean sort of who controls your data, but also like who controls who gets the internet. Whether you're planning a big tech event, launching a new campaign, Thank you You can customize it all your logo your message your look and many items come with no setup charge to help you save And if you really watching the bottom line you find standout choices at every price point so you can make a real impact while staying on budget Plus, you'll get expert help, fast turnaround times, and their 360-degree guarantee, so you can be 4imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint, 4certain. Support for the show comes from public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions. Buying the dip, manually sweeping idle cash, putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S&P 500. or if my cash balance goes above $20,000, move the excess into my direct index. You approve the workflow and your agent handles the rest. Monitoring the market, watching for your conditions, and executing your strategies exactly as defined. An investing platform driven by your intent not just your clicks You can also get full read and write access to your account via the public API Go to public slash market and fund your account in five minutes or less That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., Member FINRA, and SIPC. Advisory services by Public Advisors, LLC, SEC Registered Advisor. Complete disclosures available at public.com slash disclosures. So there's a lot of noise about AI, but time's too tight for more promises. So let's talk about results. At IBM, we work with our employees to integrate technology right into the systems they need. Now, a global workforce of 300,000 can use AI to fill their HR questions, resolving 94% of common questions. Not noise. Proof of how we can help companies get smarter by putting AI where it actually pays off. Deep in the work that moves the business. Let's create smarter business. IBM. When the infrastructure of the internet was first starting to be built out, particularly in the 1990s, there was a popular parallel to the railway booms, etc. And it's sort of similar economics and connecting locations and all that Rail is kind of a natural monopoly You know if you want to have a rail between New York and Chicago you don need five separate tracks owned by separate companies all right next to each other It's sort of wasteful, etc. Wait, let me just say as a regular Amtrak commuter, I do wish there were multiple tracks. You wish there were competition, but there's never going to be, right? That'll just never happen. There aren't going to be four tracks between a random town in Connecticut and New York. One can dream. One can dream. To what degree do the economics of this replicate? how much redundancy is there? How many different cables are there in some cases along exactly the same route? To what degree does the natural monopoly effect either replicate or not? Well, there's a considerable amount of redundancy built in because I think the deal here is this, even if Google owns, for example, one transatlantic cable, and even if it, let's assume hypothetically, it funnels only Google-based data through that cable. So let's assume funnels Gmail and Google Meets and all this other stuff. Only Google data is going through that cable. This never happens, but we'll let it ride for now. It is in Google's best interests to buy redundancy on another cable because that cable might go down. And in a similar way, it is in Meta's best interest to buy redundancy on a Google. [20:00] cable just in case a meta cable goes down. So there's a lot of redundancy built into the system. There's between 500 and 550 undersea cables around the world. Many of them are sort of clustered around these, not surprisingly, these big areas of economic activity. So Western Europe to the eastern seaboard of the US, Southeast Asia, China and Southeast Asia, the Gulf and then India. So these are heavily trafficked routes and there's a lot of cables that traffic them for the simple reason that redundancy is important. We can think about the fact that every year, roughly 100 cables get cut around the world. Most of these are accidents. There might be a ship that throws its anchor overboard and cuts a cable by accident, or it might be that a fishing boat is trawling the seabed and it hooks a cable and snags it and cuts it. So think about 100 cuts every year. And yet we don't experience nearly that order of internet outages, right? And part of the reason is, of course, we have redundancy in the form of land cables, which I haven't really talked about in this book, but there's cables coming to the US from Canada and so on. So that takes up a share of the traffic but also there redundancy in the marine cable ecosystem itself So like if one cable gets cut there enough cables crossing the Atlantic for example to take up the burden of the data that was severed Actually, this reminds me, the other thing I wanted to ask you is how much does geography play into these decisions? Because again, going back to that chart, they very much resemble shipping lanes. So getting from point A to point B, presumably as fast and efficiently as possible. But also one thing I learned in researching for this episode is that Egypt apparently is a big, I don't want to say choke point, but like a major center for fiber optic cables in much the same way that it's a major center for containers shipping through the Suez Canal. So how much is just pure basic geography actually inform the decisions of where these cables get laid? Oh, it's a big deal. I mean, first as undersea geography, right? I mean, the topography of the ocean floor is obviously not flat, flat, flat. There are crevasses and gorges and there's currents that go one way or another. There's like steep lips of rocky cliffs. And so you have to kind of avoid all of this. You know Tonga which I mentioned earlier has the unfortunate situation of sitting in that Pacific ring of fire So there an underwater volcano not far away and that liable to break its cables And it has in the past even before the one that I talk about So undersea geology geography big deal But even overseas geography is a massive deal So for example, if you want to come from Asia to Europe, as with a ship, it's much cheaper to come through the Suez Canal rather than to go around South Africa. And, you know, in the Neil Stephenson essay I talked about, he makes a big deal out of this. It's a really elegant way. He goes to Alexandria, where a bunch of cables are sort of crisscrossing over from Asia to Europe and vice versa. And he talks about how Alexandria is at that point kind of like a storehouse or a hub of information, very similar to how it was back in the day when the library of Alexandria was still in place. So there's this parallel with ancient history. But you bring up Egypt, and Egypt is a really good example. The other good example is something that's very much in the news now, which is the Strait of Hormuz. And just as it's very convenient for ships to take that path in, so it is for cables to take that path in. And the Red Sea and the Strait of Hormuz are both choke points, as you say again in which if a malign actor sort of wanted to really cripple the global internet they could go down there use your lobster drones that you talked about and clip every single cable that runs through the ocean floor over there Wait can I just ask on that So when I look at the map of subsea cables going across the Strait of Hormuz it's like Saudi Arabia connecting to, I guess, Iran. Can they cripple the global Internet by snapping those cables in particular, or would it just cripple the connectivity between the countries that are clustered around there? Well, I mean, it would definitely impede the countries clustered around there. But we also have to remember that the Internet doesn't function like a highway in the sense that the fastest route between two points is not necessarily the shortest route. So, for example, if I'm sitting in London and I'm sort of pinging a server in, let's say, Portugal, it may well be that at that particular moment and this router is making the decision for me, the shortest route is to France and then onward to Portugal by land rather than through sea throughout. I see. It may also be that I live in Saudi Arabia, but I'm a Gmail user and my Gmail data is being stored somewhere else. It's being stored maybe in Western Europe or it's being stored even in the US. You don't know which server your data lives on at all. And so when you cut a bunch of cables that service that particular part of the world, [25:00] Also, essentially forcing the rest of the Internet to reroute itself constantly until these cables are fixed. How does cable repair work? Oh, man, it's so old school. I mean, I was kind of shocked at this. I thought there would be underwater, unmanned underwater vehicles and so on and so forth. But actually, a lot of it is just exactly how telegraph lines were fixed back in the day, which is that you send a ship out to where you think the cut is. And obviously now our sense of where a broken cable might lie is much, much sharper than it used to be in the 1800s. But once you get there, you throw a grapnel hook overboard and you kind of drag it along the seafloor and you hope that it snags the cable. And sometimes you have a bite and you pull it up and it's caught something else entirely. And you chuck it back into the ocean and you try again. And so this is essentially how all ships do repairs these days. And of course, once you get the cable on board, it then jumps back into extremely sophisticated gear. So there's a lab on board the ship that is built to splice sort of glass fiber optic cables together. It stabilized in a even in a very rocky sea it can sort of work with astounding levels of stability There a clean room And so you kind of do all of this stuff and then you carefully lay it back into the ocean and the exact coordinates where it used to be And you test it and you test it again and again until you make sure that it works So we know that cables can be cut mostly by accident, but occasionally by bad actors. But as you mentioned earlier, there are redundancies built into the system. What are redundancies that I guess we're not as attuned to at the moment? Like, what are the weaknesses that we should be really thinking about? Well, I mean, one of the weaknesses, this is not a weakness from the internet's point of view, maybe, although some people might beg to defer. But I think this earlier point that I raised, which is the dominance of four essentially American tech companies in the cable-laying world is a weakness in one way. It kind of shrinks the internet because the internet then depends on where these companies want to lay cables and people rely on them for who's going to get served by the internet and who isn't. There's an example that I cite in the book about the longest cable laid anywhere in the world. And this is a cable called 2Africa and it's funded and owned, I think almost entirely by Meta And Meta therefore I think makes the decisions of well which West African countries should we land in Now West African countries and Africa in general they very underserved by the Internet I mean, this is sort of an Internet deficient population. And for any economy today, I think the Internet is sort of it's a prerequisite. You can't develop and grow without being online as much as everybody else is in the West. And so for the government of West Africa, of Cote d'Ivoire, for example, this presents a real dilemma. On the one hand, do they welcome Meta in and say, you know, of course, feel free to land a cable on our shores. But we have these kinds of conditions that European countries have. We want our citizens' data to be stored on servers on our soil. And this is something that European countries have been able to demand and get quite often. But unfortunately, the Cote d'Ivoire government has much less leverage than any Western European nation. And so, for example, if we were to tell Meta that, it runs the risk of Meta saying, well, that's fine. Maybe we don't want to land in this country after all and loop somewhere else instead. And so then there's a real dilemma. I mean do they sort of look out for the privacy and data security of their citizens or do they give their citizens expanded Internet access which is really what they craving And I think this is a weakness It gives the lie to the old sort of utopian model of a free open Internet And as in many other ways, which we talk about all the time these days, I think this also actually sort of closes off the Internet in some sense and kind of suffocates certain people who want to get online. Joe, I was trying to think of a really remote place in the world that might not have fiber optic cable yet. And so I looked up Kiribos and it turns out there's a cable that's supposed to have become operational funded by the US, Australia and Japan in 2025. But I can't see anything about whether it actually did start working. There you go. There must be some Chinese company that's competing with Meta and all these going to the developing world in Africa and LATM and saying, you don't have to trust these big tech giants. We'll do it. We'll do it cheaper, et cetera. That must exist, right? There is exactly one. Well, there's two, I think. But the main one is called HMN, which is a spinoff of Huawei. And HMN has been in the cable business for a while now, like I think more than a decade. And HMN doesn't really own cables outright. HMN is kind of like these two or three big companies, including Subcom in the U.S., that lay cables. [30:00] go out there and they lay the cable for a corporate customer. And so you could easily go to them and say, why don't you lay our cable instead? It'll be sort of cheaper than these American companies that do it. The problem here is that as for the last six or seven years, Chinese tech companies have been on sanctions lists around the world after the first Trump administration put them there. So to go out and engage HMN and to ask Huawei to come and lay a cable is to sort of risk going against that sanctions list and therefore risk all kinds of punitive measures being enforced against you. There used to be a time that the idyllic days about eight or nine years ago, when Western companies and Chinese companies would come together in consortiums and lay cables jointly, they would kind of bear part of the burden and they would figure out who was going to do the actual laying and how much bandwidth people would get. Two or three of these projects have actually been frozen. In one case, the cable has actually been laid, but nobody can switch it because the US government doesn't want that to happen because of Chinese companies involved. In two other cases, I think the mapping was done, the funds were committed, and then they had to kind of pull everything back. There hasn been a new cable laid between China and the US direct in years in years even though traffic between these two countries has never been greater So you have to route everything now through the Philippines or through Singapore It's like a very, very long way around to go from Chinese East Coast to the West Coast of the U.S. So none of this is, of course, efficient in any way or desirable in any way. But even more than that, I think people in the industry start to worry that something like a kind of bifurcation of the internet is going to start setting in. By which I mean that if there's a particular part of the world that needs cabling, and the Americans want to cable it but don't want Chinese involvement, and the Chinese want to cable it as well, there will essentially be two parallel cable systems running in that part of the world. Now, this isn't parallel in the sense of redundancy. This is parallel in the sense of superfluity. You don't need both of them, and yet you have both of them. And so what happens to the internet there? Does it get fractured and split along these two cable systems? What then happens to the promise of the Internet, which, as we said, was conceived as a free and open access kind of system to link the world? And so I think there are people in the industry who talk to me quite a bit about their anxieties for just how the Internet might be bifurcated. Thank you can find the right fit for any audience purpose or budget you can customize it all your logo your message your look and many items come with no setup charge to help you save and if you're really watching the bottom line you'll find standout choices at every price point so you can make a real impact while staying on budget plus you'll get expert help fast turnaround times and their 360 degree guarantee so you can be for imprint certain your order will arrive on time and look exactly right. Whatever your goal, 4imprint makes it easy to find your perfect promo match. Explore the possibilities today at 4imprint.com. 4imprint, for certain. Support for the show comes from public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions buying the dip manually sweeping idle cash putting on a hedge On public you can now create AI agents that handle all these tasks on your behalf Just describe what you want to do in plain English Like, if the VIX hits 25, buy a put option on the S&P 500. Or, if my cash balance goes above $20,000, move the excess into my direct index. You approve of the workflow and your agent handles the rest. Monitoring the market, watching for your conditions, and executing your strategies exactly as defined. An investing platform driven by your intent, not just your clicks. You can also get full read and write access to your account via the public API. Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., Member FINRA and SIPC. Advisory services by Public Advisors, LLC, SEC registered advisor. Complete disclosures available at public.com slash disclosures. So there's a lot of noise about AI, but time's too tight for more promises. So let's talk about results. At IBM, we work with our employees to integrate technology right into the systems they need. Now, a global workforce of 300,000 can use AI to fill their HR questions, resolving 94% [35:00] of common questions. Not noise. Proof of how we can help companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business. IBM. It came up that most of the damage that happens to undersea cables is probably by accident, fishermen, whatever. But is there a measurable increase in intentional sabotage of cables that we've seen? I mean, people within the industry seem to think so. And I think governments seem to think so as well. And so we've seen an amping up by many degrees of the kinds of protective measures that governments and companies accord to these cables. So the UK, where I live right now, a few years ago, they announced that they would have two military naval vessels permanently on patrol around the island to protect not just undersea data cables, but also power cables and oil and gas pipelines. The Baltic nations and the Scandinavian nations have their Coast Guard patrols on higher alert than ever before because they think Russian ships are out there to cut these cables in acts of what is known as gray zone warfare. The Taiwanese government is paranoid that China will eventually you know in some kind of act of war cut every single one of the 15 internet cables coming from overseas that land on Taiwanese shores So I think there like a real anxiety surrounding this particular act of malicious damage. This, of course, goes along with the fact that there's plenty of other anxiety about geopolitical conflict in this day and age. But I think countries and governments are acutely aware that this is one extremely effective way of crippling economies that a malicious actor could undertake at virtually no or little expense. So I'm looking at a headline right now that says Saudi Arabia, Qatar and the UAE are financing competing data corridors through Syria, Iraq and East Africa to try to bypass the Strait of Hormuz to avoid the problems we've been talking about. What are the things that governments or private companies can actually do if they're worried about bad actors out there who might be attacking this vital infrastructure? Well, one thing to do is to build out more redundancy. And I don't mean just lay more cables. If you look at the map of the telegraph cables of the late 1800s, and you look at the current map of undersea cables you see that they kind of similar I mean they almost the same except for this big new thicket of undersea data cables that has come up in East Asia and Southeast Asia And part of the reason for that is out of habit or inertia or reasons of cost efficiency cables still land in pretty much the same places that they did 100 years ago So I'm from India, for example, and many, if not all, of the undersea data cables that land in India today land in Mumbai on the West Coast or Chennai on the East Coast. And that's exactly where the telegraph cables used to land a century ago. I asked many people as to why this is. And, you know, it's just easier. You have the data centers there already. There's economies of scale. So all you have to do is just get an extra cable on there and you have the rest of the infrastructure ready to go. But maybe your solution lies in finding other places to land, other places in which this data infrastructure can be protected. That's one option. The other option is exactly what you said right now, which is to find alternate routes. But this is not easy either. I mean, the reason, for example, that all these years cable companies continued laying through the Strait of Hormuz and didn't go through Syria is because Syria was a hotbed for conflict. I mean it very difficult to imagine laying a cable through Syria circa eight years ago or nine years ago or laying a cable through Iraq circa 25 years ago or indeed 30 years ago These were places where geopolitical sort of turbulence made it very difficult to ensure the security of the cable. But yet, as we see now in the headline you read out, people are being forced to reconsider these other routes, which I think is a good thing as well. You mentioned that it's just not plausible right now to replace all of this infrastructure with basically satellite internet the capacity doesn't exist for all kinds of reasons you know the elon i was gonna say the elon musks of the world but there's actually one elon musk but you know he's sort of in a category of his own in many respects but like is there a vision other people talking about the prospect that one day the entire internet could truly be wireless i mean they find ways 5g they find ways to optimize wireless connectivity that are significantly better than they were five years ago, certainly 20 years ago, etc. Is anyone talking about a post-cable future? No, I don't think anybody is. And the reason for that is however much satellites have improved, and they have a lot. So, for example, in the 16 months in which it didn't have cabled internet, one of these outlying islands of Tonga, [40:00] This relied on Starlink throughout. And that was very useful. But the problem here is that however much satellites improve, our appetite, our sheer thirst for data is just going to improve exponentially. We're seeing that now with the AI boom and the fact that all of this training data is also sitting on servers in one place and we're calling it up on Clawed and ChatGPT and other kind of engines on our laptop. And so this is a new kind of use for data and a new kind of data flow that wouldn't have existed five years ago. So we've clearly already seen this explosion of additional data burdens on the cabling industry. And so I think, realistically speaking, nobody thinks that we can please replace all our data infrastructure with satellites. I think cables are here for a good long while yet. This was going to be my next question, actually. So your book came out, I think, just in October of last year. So still relatively new. So how does the rise of AI, the hyperscalers, the data centers actually change the trajectory of the subsea cable world? I don think it changes the trajectory so much It just kind of accelerates it I mean there was already as I mentioned earlier some of these big tech companies that were laying cables of their own All those big tech companies are now engaged in their own kind of AI efforts and often very big large scale ones at that And so I think like the need for these cables and the need to lay more cables and at a faster clip is only going to increase because I think you're going to have more AI training data centers out there. You want to have more flows of data. And so people were kind of worried for a little while, maybe about five or six years ago, people were talking about the fact that maybe there was a glut of cables out there right now, and that maybe we didn't need to lay any more new cables for the foreseeable future because we were kind of set. But the AI boom has changed all that. It is also true that the cabling industry follows a similar trajectory to the world economy. When there's a recession, it goes down, people stop laying cables because it's very expensive to finance as a project. After the dot-com boom, for example, in early 2000 and the dot-com bust, there was a period of stasis that lasted for almost like five years. There was all this extra cabling that had been ordered up, but nobody was laying it because it was just too expensive. So I think part of it is cyclical and it will follow the trajectory of the economy itself But I think as far as the AI revolution or the AI boom goes as long as that persists I think we still need to see more and more cables being laid just to be able to just to make this entire project possible One last question I have. So one of the reasons why cable going back to the old days exists is for finance specifically and having information about trading in one place and another at roughly the same time making markets more efficient. Another thing is corporations, they don't use the public Internet for a lot of their communications. They use private Internet, which are faster and more secure and safer. Are there examples of like, I don't know, JP Morgan, they have offices in London and they have offices in New York wanting to have purely dedicated cables that is not the public Internet, but is essentially a very large scale expansion of their private Internet? no I don't think that exists yet I mean there have been instances in the past I think of like financiers deciding that they want to lay their own overland cables yeah I can't remember which book it was I think maybe it was rocket boys but it was a Michael Lewis book that oh about flash boys flash boys yeah flash boys that right that opens with the laying of a cable overland not overseas not under sea but you know that does exist but I think the only sort of big entities that decide that they need their own cables and they going to lay them and exclude everybody else from them are the big militaries of the world. So the U.S. military has its own cables. The Chinese military has its own cables. They have their own contractors for that. And none of the public Internet travels on them whatsoever. What's your favorite subsea cable? oh man there's a really short one that almost shouldn't exist because it's so short but i feel like it's so cute because it does exist it's i don't know maybe 14 miles long i'm looking this up in my book now and it goes between the isle of man and northern ireland it's 41 miles so it's not super short but it just feels like it's one of those cables that that was laid only because it demonstrated just how cost-efficient these technologies were. All right, Samantha Supermanium, we're going to have to leave it there. But thank you so much for coming on All Thoughts. That was super interesting. Thank you so much. A lot of fun. So, Jo, that was fascinating. [45:00] It's a lot to kind of pick over in that discussion. I'm very interested in how we, I mean, it's sort of a parallel of the global economy, isn't it? Where you used to have consortiums who would work together to build these things. And now it's just sort of one of four very large tech companies that are doing most of it. The other thing that I was thinking about is just, I didn't realize the militaries had their own cables. But again, it makes a lot of sense. Yeah, no, I love that topic. Like I said, I still think it's one of the great marvels. marvels like there's a lot of miles yeah except we can't actually marvel at it because it's underneath miles and miles of the ocean well i can't go and look at it i've like problem i've pulled up the bbc website for example and so i kind of can appreciate that map the no no i mean i'm literally i've been to websites from other countries so i sort of marvel at the speed with which they get to me it really just blows my mind the other thing that's just from a technical standpoint, not an economic or geopolitical standpoint, the fact that you can send a beam of light through a thing and it bounces around this tube and then it winds up in exactly a configuration that can be recombinated in exactly where it starts that never ceases to blow my mind that fact But the other thing I was thinking about was samantha mentioned that there was a period where no new subsea cable was being laid yeah partially because it was a recession and people don't want to spend a lot of money but also because it it reminded me a lot of electricity in the power industry where like the loads needed by the population start to flatline and so we don't really invest in it for a while and then suddenly there's a technological advance like AI and suddenly we need to we're all thinking about our electricity needs going up. Feels very much like that. There's the thing about the dot com era, which is that like in 2002, it was like, oh, we're never going to use all this cable. And then by like three years later, they're like building it again. But like the dot com era looks much less crazy in retrospect because they were all completely correct about how big the Internet was going to be and how much demand was just it ended up just being a slight timing mismatch more than anything else. Technology can be underhyped and overvalued at a single point in time. That's right. That's what I keep saying. All right. Shall we leave it there? Let's leave it there. This has been another episode of the All Thoughts Podcast. I'm Tracy Allaway. You can follow me at Tracy Allaway. And I'm Jill Weisenthal. You can follow me at The Stalwart Follow our guest Samantha Supermanian He at Samantha underscore S Follow our producers Carmen Rodriguez at Carmen Armand Dashiell Bennett at Dashbot and Kale Brooks at Kale Brooks And for more OddLots content, go to Bloomberg.com slash OddLots for the daily newsletter and all of our episodes. You can chat about all of these topics 24-7 in our Discord, discord.gg slash OddLots. And if you enjoyed this conversation, if you like it when we talk about subsea cables and lobster drones, then please leave us a positive review on your favorite podcast platform. And remember, if you are a Bloomberg subscriber, you can listen to all of our episodes absolutely ad-free. All you need to do is find the Bloomberg channel on Apple Podcasts and follow the instructions there. Thanks for listening. We be right back giveaways client gifts or team gear 4imprint has the promo products to match With thousands of options from apparel and drinkware to tech and totes it easy to find the right fit for your brand and budget with standout choices at every price point. And with their 360-degree guarantee, you can be 4imprint certain your order will show up just right, right on time. Explore more at 4imprint.com. 4imprint. 4certain. When you're running a business, the best days are the ones where priorities stay on track. For midsize and large companies, risk can affect multiple parts of the organization at once, from property and liability to cyber and regulatory challenges. At that level, managing risk becomes an ongoing discipline. At the Hartford, the focus is on helping businesses manage risk before it turns into something more disruptive. And when losses do happen, that work is paired with insurance coverage shaped by years of underwriting, risk engineering, and claims experience. Learn more at thehartford.com slash risk mitigation. Policies provided by Hartford Fire Insurance Company and its property and casualty affiliates, Hartford, Connecticut. The U.S. Virgin Islands invites you to explore the stunning islands of St. Thomas, St. Croix and St. John. USVI is truly a one of a kind destination with mouthwatering food, the warmest hospitality from the locals, vibrant culture and prestige. [50:00] teaches usvi has everything you're looking for in a getaway and the best part no passport is required when traveling from the u.s go to visit usvi.com to start planning your next vacation that's visit usvi.com usvi naturally in rhythm