Mackinder, Redrawn 麦金德，重画

In April 1904 Halford Mackinder stood up at the Royal Geographical Society and read out a paper that has haunted foreign-policy language for 122 years: "Who rules East Europe commands the Heartland; who rules the Heartland commands the World-Island; who rules the World-Island commands the world." The Heartland was the inland mass of Eurasia — beyond the reach of the sea empires, newly mobile by rail. He drew it as three concentric rings on a Mercator map: a pivot, an inner crescent, an outer crescent.

The thesis assumes that geography means territory. In 2026 it largely doesn't. The locks of power today aren't regions; they are chokepoints — single nodes small enough to fit on one campus, one strait, one mine, whose failure reorders an entire sector. TSMC's Hsinchu fabs are a few square kilometres and produce ninety per cent of the world's leading-edge logic. The Strait of Hormuz is thirty-nine kilometres wide and carries one-fifth of seaborne oil. Bayan Obo is one iron-mine complex in Inner Mongolia and refines roughly seventy per cent of the world's rare earths. ASML in Veldhoven makes 100% of the EUV lithography machines that every advanced fab on Earth depends on. None of these is "territory" in Mackinder's sense. Each is a node.

Read the dots on the globe and one pattern surfaces immediately: nearly every binding chokepoint of 2026 sits on Spykman's Rimland — the coastal arc of Eurasia — not on Mackinder's Heartland. Hormuz, Malacca, the Red Sea cables, TSMC, ASML, Korea's fabs, Suez. The Heartland still holds the geological inputs that the Rimland's nodes need (Bayan Obo, Russian gas, the Belt-and-Road land hub at Khorgos), but inputs are not decision points. Chokepoints are. Spykman saw this in 1942; Mackinder never quite did.

Geography hasn't gone away. The map of 2026 is not flatter than Mackinder thought — it is lumpier. Power concentrates at nodes, not in zones. Whoever holds the most chokepoints commands the most. Nobody commands the world.

What this map deliberately leaves out. The smuggling and grey-export channels that connect the nodes — chip transhipment via Singapore, mineral laundering via Rwanda, dollar workarounds via stablecoins. They are real, growing, and by design unmeasurable.

1904 年 4 月，Halford Mackinder 在英国皇家地理学会宣读了一篇论文， 之后 122 年里，这段话一直笼罩着外交政策的话语： 「谁统治了东欧，便主宰心脏地带；谁主宰心脏地带，便主宰世界岛； 谁主宰世界岛，便主宰世界。」所谓心脏地带，是欧亚大陆的内陆—— 海权够不到，铁路时代让它重新拥有了机动力。麦金德把它画成 墨卡托投影上的三个同心圆：一个枢纽、一个内新月、一个外新月。

这套论断假设地理就是领土。今天大体上不是这样了。2026 年， 权力的关节不在区域，而在咽喉点——小到一个园区、一道海峡、 一座矿，它一旦失灵，就能搅翻整个行业。台积电在新竹的几平方公里厂区， 做出全球 90% 的最先进逻辑芯片。霍尔木兹海峡只有 39 公里宽， 五分之一的海运原油从这里通过。白云鄂博是内蒙古的一座铁矿， 精炼着全球大约 70% 的稀土。荷兰维尔德霍芬的 ASML 是地球上唯一能造 EUV 光刻机的公司——所有先进晶圆厂都得指望它。这些都不是麦金德 所说的"领土"，它们是节点。

把这些圆点放回地球上看，一个规律立刻就显出来：2026 年几乎所有起作用的 咽喉点，都落在 Spykman 的边缘地带——欧亚大陆的海岸弧——上， 而不是麦金德的心脏地带。霍尔木兹、马六甲、红海光缆、台积电、ASML、 韩国晶圆厂、苏伊士。心脏地带依然握着边缘地带节点离不开的那些地质资源 （白云鄂博、俄罗斯天然气、霍尔果斯的一带一路陆港），但原料供给不是 决策权，咽喉点才是。Spykman 1942 年就看清了这件事，麦金德始终没看见。

地理并没有失效。2026 年的世界图，不是比麦金德画的更平， 而是更斑驳——权力不再凝聚在区域里，而是凝聚在节点上。 谁握有更多咽喉点，谁号令的就更多。没有人号令世界。

这张图刻意没画的东西。把各节点串起来的走私与灰色出口通道—— 经新加坡的芯片转口、经卢旺达的矿产洗白、绕开美元的稳定币结算。 它们真实存在、规模在长，且本就被设计成无法被精确计量。