An extension to The Race for the Bomb, our ten-part focus series. Part 7 told the engineering story of the B-29 — the pressurised cabin, the burning engines, the Battle of Kansas. This piece asks the question that story leaves hanging: how did America build nearly 4,000 of the most complex aircraft ever made in four years — and why can nothing like it be done today?

Willow Run, Michigan, spring 1944. A four-engined B-24 Liberator — 1.2 million parts, more than 300,000 rivets — rolls off Ford’s mile-long assembly line. Sixty-three minutes later, another one follows. Then another. Twenty-four hours a day, seven days a week: in April 1944 alone, this single building delivered 428 heavy bombers.

Now the modern mirror. In 2025 — a record year — the three F-35 plants on three continents delivered 191 fighters between them: roughly one every 46 hours. America’s new stealth bomber, the B-21 Raider, is believed to be built at seven to eight aircraft per year. In 1944, the United States produced 96,270 aircraft. In one year.

These numbers are not a story about people being tougher in the old days. They are a story about what an economy chooses to be — and the B-29 Superfortress, the bomber at the heart of our series, is the sharpest lens for looking at it, because it was simultaneously the most complex machine mass-produced to that point in history and one of the fastest big programs ever run.

The Plane That Proved It Could Be Done

Recall the timeline from Part 7. The Army signed the XB-29 prototype contract in August 1940 and ordered 250 production aircraft in May 1941 — sixteen months before the prototype ever flew. First flight came in September 1942; the first combat mission, against Bangkok, in June 1944. From a drawing to combat over Asia: 46 months, for a pressurised, computer-gunned aircraft at the absolute edge of what 1940s engineering could do, at a program cost of three billion dollars — more than the atomic bomb it would carry.

And it was built everywhere at once. Four final-assembly plants ran in parallel: Boeing’s lines at Wichita (1,620 aircraft) and Renton (1,119), Bell’s at Marietta, Georgia (668), and the Glenn L. Martin plant at Omaha (531) — the line that built Enola Gay and Bockscar, which is where this article joins hands with the rest of the series. Four companies, four states, one design. Today the United States has five major defence primes in total.

None of it was clean. As Part 7 recounted, of the first 97 B-29s built at Wichita, sixteen could fly, and the Battle of Kansas — 1,200 technicians resoldering 586,000 connections outdoors in a Kansas winter — was what turned a broken fleet into a combat force in ten weeks. That is the number modern procurement should envy most: not the production rate, but the recovery rate.

The Arsenal

The B-29 was the extreme case of a machine the whole economy had become. When President Roosevelt told Congress in May 1940 that he wanted America able to build 50,000 aircraft a year, the country was making fewer than 3,000 annually and the number sounded like rhetoric. By 1944 it looked conservative: 46,907 aircraft in 1942, 84,853 in 1943, 96,270 in 1944 — roughly 300,000 by the war’s end, two-thirds of all Allied military equipment coming from American factories.

The muscle behind the slogan was human and improvised. By late 1943 some 475,000 women worked in American aircraft plants; Willow Run alone employed 42,500 people and delivered 6,972 complete B-24s plus 1,893 knock-down kits. The shipyards ran the same play — the average Liberty ship build fell from 355 days to 41, and one yard famously launched a complete ship four days and fifteen hours after laying the keel, as a stunt, to prove the ceiling had not yet been found.

The Modern Mirror

Now set the same stopwatch on the present. The B-21 Raider — by every account a superbly run program, on cost and broadly on schedule — received its development contract in October 2015 and first flew in November 2023: eight years, against the B-29’s 25 months. It is expected to reach operational service in the 2026–27 timeframe at a reported rate of seven to eight aircraft per year, toward a planned fleet of at least a hundred. At those figures, building the B-29 fleet of 3,970 aircraft would take roughly five centuries.

The pattern repeats across every category. The F-35 took fourteen years from contract to initial operational capability, and its record 191 deliveries in 2025 came from a global production system 25 years into its life. The B-2 before it: 21 aircraft at an average of 2.13 billion dollars each. Even the KC-46 tanker — a derivative of a 1980s airliner — needed eight years from contract to first delivery. On the civilian side, Boeing’s four 737 lines are currently climbing from 42 toward 47 aircraft a month and treating that as an industrial achievement; Renton, one of those same factories, once turned out 1,119 B-29s in under three years.

And yet the appetite for more is suddenly real. In 2026 the Air Force announced expanded B-21 production capacity, Congress added billions to accelerate it, and Northrop Grumman is investing 2.5 billion dollars of its own money in the line. The Pentagon’s leadership now talks openly of wanting substantially more aircraft.

Why We Can’t Just Do It Again

So why can’t a nation that built 96,270 aircraft in 1944 build 96 bombers in a decade without heroic effort? Five structural reasons, none of them laziness.

First, the machines changed species. A B-29 carried five analog fire-control computers and zero lines of software; the F-35 flies on roughly 24 million lines of code, and most of a modern program’s schedule is not bending metal but writing, testing and certifying software and sensors. Second, the industry consolidated: at the Pentagon’s famous 1993 “Last Supper” dinner, contractors were told the budget could not sustain them all, and 51 major defence firms became five within four years. The four-plants-one-bomber model of 1944 has no modern equivalent to activate. Third, the contracts inverted: wartime production ran cost-plus, with the state absorbing risk; today’s fixed-price programs punish speed — Boeing has lost roughly five billion dollars on the KC-46, Northrop over two billion on early B-21 lots. Fourth, the risk tolerance is gone, and mostly for good reasons: 1944 accepted dead test pilots, dead crews and sixteen flyable aircraft out of 97 as the cost of momentum. No democracy at peace accepts that arithmetic. And fifth, there is no total mobilisation: no draft of the whole labour force, no conversion of car plants, no 475,000 recruits to the rivet gun.

The B-29 itself is the honest summary of the trade. Its speed killed people — Eddie Allen among them — and its program style would be criminal negligence in peacetime. But it also proved what an industrial democracy can do when it decides survival depends on it: from nothing to 4,000 examples of the most advanced aircraft on Earth, faster than a modern program moves from contract to first flight.

The uncomfortable question, which 2026’s budget lines are quietly asking, is whether the avalanche can be summoned again in time if it is ever needed — or whether Knudsen’s machine was a one-time invention of a mobilised century. The optimists point to the B-21 line accelerating; the pessimists point to the calendar. The Willow Run clock — sixty-three minutes — is still the record, and nobody is currently trying to beat it.

Sources: Wikipedia (B-29, Willow Run, US aircraft production during WWII, B-21 Raider, F-35, B-2, KC-46, Last Supper), Air & Space Forces Magazine, Breaking Defense, Leeham News, The American Presidency Project, Library of Congress.