THE IRON DAIMYŌ (1850–2026) — POWER OF THE GREAT RAILROADS, IN ROUTE-MILES

Why the Great Northern was so powerful. Three structural advantages, compounding for a generation. (1) No land grants, no debt bomb: every other transcontinental was built on federal land grants and bond subsidies — which meant building fast, badly, and ahead of any traffic, then drowning in interest. James J. Hill built the Great Northern with private capital, extending only as fast as paying freight followed, and reached Seattle in 1893 — the exact year the subsidized roads collapsed around him. It is the only transcontinental that never passed through receivership. (2) The best engineering in the West: Hill personally scouted grades; his engineers found Marias Pass, the lowest crossing of the northern Rockies — and low grades meant every train ran cheaper than a rival's, forever. "What we want," Hill said, "is the best possible line, shortest distance, lowest grades and least curvature." (3) He grew his own traffic: Hill imported farmers by the trainload, ran demonstration farms, gave away breeding stock — colonizing his own service territory the way the Hōjō drew peasants into the Kantō. The Empire Builder didn't find an economy to serve; he built one.

Why the house endured while the flashier lords fell. Strategy: Hill fought defensive, regional wars — he took the Northern Pacific out of Morgan's reorganization and the Burlington in 1901, but never speculated, never built ahead of revenue, never bet the road. When Harriman tried to take the NP from him in 1901, the fight cornered the stock at $1,000 a share and nearly crashed Wall Street — and Hill still won. Roosevelt's trust-busters broke up the resulting holding company in 1904, but the three roads waited, aligned, for 66 years.

The consummation. In 1970 the merger Roosevelt blocked was finally allowed: GN + NP + Burlington became Burlington Northern, and in 1995 BN took the Santa Fe. Then came the greatest tribute of all: Berkshire agreed in 2009 to take BNSF private in a transaction valued at about $44 billion including $10 billion of BNSF debt, then the largest acquisition in Berkshire history. Buffett called the $34 billion investment an "all-in wager on the economic future of the United States." Hill's patient empire outlasted every rival that ever mocked its caution.

Nationalizing the northeastern railroads looked like the industry's obituary until you see what came back out. It was a forced acquisition, not a rescue anyone wanted: when Penn Central failed in 1970 — 872 days after the PRR–NYC merger, the largest bankruptcy in American history to that date — the Erie Lackawanna, Reading, Lehigh Valley, Central of New Jersey and Lehigh & Hudson River were also bankrupt. In 1976 Washington transferred the viable portions of six failed systems into one government-owned railroad, Conrail. The predecessor maps totaled roughly 34,000 route-miles before the Final System Plan selected and rationalized the operating network; subsequent federal support reached about $7 billion.

From Congress's chair, the logic was triage: (1) the Northeast could not physically function without these rails; (2) no private buyer would touch them under 1970s regulation, where the ICC still set rates as if it were 1910 and railroads were monopolies rather than roadkill; (3) so the state would hold the land until the rules changed. Then the rules changed: the Staggers Act of 1980 freed rates and let railroads shed unprofitable line — and Conrail, under Stanley Crane, cut its map nearly in half, went profitable by 1981, and in 1987 was sold back to the public in what was then the largest IPO in American history.

The regranting was the real Sekigahara. In 1997–99, CSX and Norfolk Southern fought a $10 billion bidding war for the resurrected system and, unable to defeat each other, partitioned it — NS taking 58%, CSX 42% — the largest single transfer of railroad territory since the land-grant era. Roughly 11% of the national network, confiscated by the state from six dead houses, was held for a generation and then regranted to the two eastern victors, completing the map: four great systems, two per side of the Mississippi, exactly the concentration the ICC had spent 70 years preventing. The nationalization everyone called socialism turned out to be the escrow account of the final consolidation.

How the mileage estimates were made (Fermi method): Anchors are ICC/AAR route-mile statistics and company reports: the national network grew from ≈9,000 miles (1850) to its all-time peak of ≈254,000 miles in 1916, then shrank by nearly half to roughly 140,000 today. Documented checkpoints keep the bands honest: the Pennsylvania and New York Central each ran ≈10,000-mile systems at midcentury; Penn Central entered bankruptcy with ≈20,000 miles; Conrail inherited selected portions of six bankrupt systems and was ultimately partitioned on a 58% NS / 42% CSX economic basis; and BN was created at ≈23,000 miles in 1970. Today AAR describes a nearly 140,000-mile freight network, with six Class I railroads accounting for about 67% of freight mileage and roughly 615 short lines operating much of the rest. The four U.S.-headquartered giants' company route-mile figures overlap trackage rights and include some Canadian mileage, so this chart does not divide their disclosures by the AAR national total. Between checkpoints, shares are interpolated from system maps and reports; error bars are roughly ±1–2 points.

Two things that look wrong but aren't. (1) The total never shrinks — even though the network did. The stack shows share of whatever network exists each year, so the great abandonment (254,000 → roughly 140,000 miles) is invisible in the shape and visible only in the tooltip conversion. When a road dies, its share moves to another band or to gray; when a line is simply torn up, every band shrinks alike and shares are unchanged. (2) The gray band grows after 1980 — on purpose. "Other roads" is hundreds of houses: the granger, coal and New England lines of the first century, and then, after the Staggers Act let the majors shed branch lines instead of abandoning them, a deliberate new feudalism of roughly 615 short lines farming the branches. The gray band's post-1980 growth also includes the U.S. domains of Canadian National and CPKC.

The chart ends with four U.S.-headquartered giants because the newest conquest is not complete. Union Pacific and Norfolk Southern notified regulators in 2025 and filed a revised major-merger application in April 2026. On May 28, 2026, the Surface Transportation Board accepted that revised application for consideration, required more information, and held the proceeding in abeyance while an environmental-impact review is prepared. Acceptance starts the review; it is not approval. If the transaction ultimately closes, the map would stop being the symmetrical western-two/eastern-two order shown here and become the first U.S. transcontinental system under one corporate roof.