The Molt
Why dominant institutions don't eliminate their workforce to survive, they eliminate it to become something else.
Oracle posted a 95% jump in net income. Then it fired 30,000 people. This is not a story about artificial intelligence. It is a story about what dominant institutions have always done at the moment they decide to become something else and why the workforce that built them always pays for the transformation.
On the morning of March 31, 2026, up to 30,000 Oracle employees opened their laptops to find they no longer had jobs. No manager called. No HR conversation. Just an email signed “Oracle Leadership” no individual name, informing them their role had been eliminated, today was their last day, and severance details would arrive by DocuSign. System access was cut immediately.
The quarter before this email was sent, Oracle posted $6.13 billion in net income. Its contracted future revenue stood at $553 billion. The company had committed to $156 billion in AI infrastructure spending and taken on $50 billion in new debt and equity in January alone.
This was not a company in distress. This was a company at the peak of its current form, deliberately converting the workforce that built it into capital for the next one.
Almost every analysis published this week frames Oracle as an AI story a company replacing humans with machines. That is the surface. Underneath it is a pattern visible in four well-documented cases spanning 250 years, three continents, and three different economic systems. Understanding it requires going back not two years, but two centuries.
When a dominant institution reaches a technological inflection point, it does not decline - it consumes the human capital of its current form to fund the infrastructure of the next.
This is the molt. And it has one non-negotiable condition: it is always executed from strength, never from weakness. A failing institution cannot molt. It cannot borrow at scale, retain its customers through the transition, or absorb the reputational cost of eliminating its own workforce. Only a dominant institution possesses the power to convert its people into its next phase of control. That is both the logic and the brutality of the pattern.
Three cases drawn from three different centuries, three different cultures, three different scales of institutional power make the mechanism undeniable. A fourth, happening right now, makes it urgent.
The Cleveland Massacre
By 1872, Standard Oil was already the strongest refining operation in Cleveland. It had won the current form. Most refiners were hurting squeezed by overcapacity, falling margins, and volatile crude prices. Standard Oil was not. Rockefeller had already built cost advantages through vertical integration, railroad rebates, and relentless operational efficiency. His target was the next form: pipeline infrastructure. Control not of refining, but of transportation. The capital to fund it was distributed across 26 competing refinery operations. He moved to consolidate it.
In less than six weeks in what competitors called the Cleveland Massacre, Rockefeller bought 22 of his 26 Cleveland competitors, giving him control of one-fifth of the nation’s oil refining. He closed a number of the purchased refineries to eliminate excess capacity, then regulated production to match market demand, turning off the spigot of oversupply that had caused prices to collapse. He envisioned pipelines as an alternative transport system for oil and began a campaign to build and acquire them. By 1880, Standard Oil controlled 90 to 95 percent of all oil refined in the United States.
The workforce of 26 competing operations had been converted into the consolidated capital of one pipeline empire. Rockefeller called himself the industry’s saviour absorbing inefficiency, creating order from chaos. The eliminated refiners were not available to offer a different account.
He was not acting from weakness. He was acting from dominance. That distinction is everything. And twenty years later, another industrialist would prove the pattern was not an accident.
The profitable sacrifice
Andrew Carnegie was running the most profitable steel operation in America. Carnegie Steel was making record profits $4.5 million just before the 1892 confrontation, which led Carnegie himself to exclaim: “Was there ever such a business!”
Then he fired all 3,800 workers at Homestead.
Carnegie and his chairman Henry Frick were determined the union had to go. The skilled workers at Homestead enjoyed wages significantly higher than at any other mill in the country. The problem was not performance. It was structural: the Amalgamated Association of Iron and Steel Workers represented a claim on future capital capital Carnegie had already earmarked for vertical integration of ore fields, coke ovens, and transport infrastructure.
Anyone who has watched a senior team eliminated after a record quarter not because they failed, but because the budget they occupied was needed for the next strategic bet has seen a version of what Carnegie executed in 1892. The institutional logic is identical whether it operates through a 19th-century steel mill or a 21st-century reorganisation. The workforce is not being punished. It is being converted.
In the spring of 1892, Carnegie had Frick produce as much armour plate as possible before the union’s contract expired at the end of June. He then left for an extended vacation in Scotland, leaving his operations manager with explicit instructions. “We approve of anything you do,” Carnegie wrote; words he would later come to regret. On July 2, Frick fired all 3,800 workers. The state governor sent in 8,500 soldiers of the National Guard. Three hundred of the striking men were blacklisted for life.
The financial result: Carnegie Steel’s profits rose to a staggering $106 million in the nine years after Homestead.
Carnegie was in Scotland when the Pinkertons arrived. Larry Ellison was not on the earnings call when Oracle fired 30,000 people. The distance is not coincidence. It is a feature of the pattern. The leader who orders the molt is never present when it is executed. Proximity creates accountability. Distance preserves authority.
But if Rockefeller and Carnegie proved the mechanism at corporate scale, the Meiji oligarchs proved it at a scale that makes every corporate example look trivial
The elimination of a civilisation’s workforce
The most structurally complete execution of the molt in recorded history did not happen in a corporation. It happened in Japan, between 1871 and 1877, and its target was not a workforce in the ordinary sense. It was an entire social class.
The Meiji oligarchs who seized power in 1868 understood one thing clearly: Japan’s feudal, agricultural, militarily obsolete form could not survive the 19th century against Western industrial power. The institution needed to transform. And the workforce of the current form approximately 1.9 million samurai, roughly one-tenth of Japan’s population was consuming nearly 30 percent of government expenditure in hereditary stipends.
The samurai were not failing workers. They had been the operational backbone of Japan for seven centuries. They were, in every meaningful sense, the human capital of the existing state.
The molt was executed in four stages. In 1871: abolition of the feudal domains 270 semi-autonomous territories dissolved into a centralised state, eliminating the structural base of samurai authority. In 1873: universal military conscription commoners could now serve as soldiers, breaking the samurai’s exclusive claim on warfare, the function that justified their existence. In 1876: the Sword Abolition Edict samurai could no longer carry swords in public, the visible symbol of their social role stripped away. Also in 1876: compulsory commutation of hereditary stipends into government bonds. The 30 percent of national expenditure that had flowed annually to the workforce of the old form was redirected into railways, a modern conscript army, naval shipyards, industrial factories.
This is the logic that every system feudal, capitalist, communist reproduces at the point of transition. The institution does not ask whether the current workforce deserves to be kept. It asks whether the current workforce can become the next form. When the answer is no, the conversion begins. The mechanism does not require cruelty. It requires only the recognition that the current human capital is sitting in the budget line where the next infrastructure needs to be.
The resistance was both inevitable and instructive. In 1877, Saigo Takamori, the most celebrated samurai of his generation, a founding hero of the Meiji Restoration itself led over 40,000 warriors in the Satsuma Rebellion. The last armed uprising of the old order. They were crushed in eight months by the new conscript army the same army funded by the eliminated stipends. The workforce of the old form was defeated by the infrastructure built with its own capital.
Twenty-seven years after the molt began, Japan defeated China in the Sino-Japanese War. In 1905, it defeated Russia. A feudal agricultural state had become a world power in a single generation. The institution that executed the molt survived it. Saigo Takamori who had built the institution died leading the resistance of the workforce it had discarded.
China : Zhuada fangxiao - ‘grasping the large and letting the small go.’
If the pattern required only capitalism to operate, it could be dismissed as a market failure. China proved that it could not.
The largest execution of the molt in human history happened between 1997 and 2002, in China, under Premier Zhu Rongji and it carries one dimension that no corporate example can.
Since 1949, China’s state-owned enterprises had operated on the principle of the tiě fàn wǎn - the iron rice bowl: lifetime employment, guaranteed housing, cradle-to-grave welfare, administered by the state through its enterprises. The promise had been made by the founding of the People’s Republic. It had been kept for nearly half a century. By the mid-1990s, however, two-thirds of all state-owned enterprises were in debt, and the banks that lent them money were carrying billions in non-performing loans.
The institution executing the molt. The Chinese Communist Party was not weak. It had consolidated power completely. But it understood that the current form of the economy, built on lifetime employment guarantees and state-directed production, could not survive entry into the global trading system it was targeting. The iron rice bowl was the capital of the wrong form.
Zhu Rongji’s strategy was codified in four Chinese characters: zhuada fangxiao grasp the large, release the small. Tens of thousands of weak enterprises were privatised or liquidated, and stronger enterprises were restructured and listed on stock markets. Between 1995 and 2001, an estimated 34 million workers were laid off a period known as xiagang, literally “to step down from one’s post” a rupture that shattered the long-held promise of lifelong employment and state care. Independent estimates suggest Zhu shut 60,000 companies and eliminated around 40 million jobs.
The capital freed from the iron rice bowl funded WTO accession in 2001, the export manufacturing infrastructure of the 2000s, and the industrial expansion that produced the world’s second-largest economy.
The CCP broke a promise it had made to its own people for 50 years. Not because it was failing. Because keeping the promise would prevent the transition.
Forty million workers received the 1997 version of the 6 a.m. email. The institution that sent it became the dominant economic power of the 21st century.
No ideology provides immunity from this mechanism. It operates through communist party-states and capitalist corporations with identical internal logic. That is the fact that should make every founder, every board member, and every institutional strategist pay attention not because the mechanism is wrong, but because understanding it is the difference between navigating a transition and being consumed by one.
Which brings us back to the morning of March 31, 2026.
Oracle’s co-CEO Mike Sicilia stated this year that AI coding tools inside Oracle are enabling smaller engineering teams to deliver more complete solutions more quickly. This is the institutional announcement of the molt, translated into executive language. The human capital of the current form - the engineers, architects, database administrators, cloud specialists who built a $553 billion revenue backlog is being converted into AI data centre infrastructure.
The financial logic is not difficult to follow. Oracle committed to an aggressive AI infrastructure buildout requiring an estimated $156 billion in capital spending. The workforce reductions are expected to free up $8 to $10 billion in annual cash flow. The company disclosed a $2.1 billion restructuring plan in its March 2026 SEC filing, with nearly $1 billion already recorded. Employees across the United States, India, Canada, and Mexico received termination emails at approximately 6 a.m. local time, with no prior warning, no conversation with a direct manager, and immediate system lockout.
The mechanism is identical across every case examined here. A dominant institution at or near the peak of its current form. A technological inflection point requiring a different kind of capital. The workforce of the current form converted by legal manoeuvre, government edict, private army, 6 a.m. email into the infrastructure of the next.
Rockefeller converted refinery workers into pipelines. The Meiji oligarchs converted samurai stipends into railways and conscript armies. Zhu Rongji converted the iron rice bowl workforce into WTO-ready manufacturing. Oracle is converting enterprise software engineers into AI data centre capital.
The scale changes. The century changes. The culture changes. Across these four cases, the mechanism repeats.
The mechanism
The historical record on the molt has one consistent finding.
Every institution that executed it from strength survived and dominated the next form. Standard Oil controlled 90 percent of American oil. Meiji Japan became a world power in a generation. China became the world’s second-largest economy. The pattern has never failed the institution.
It has never protected the workforce that funded it.
Carnegie Steel’s profits rose to $106 million in the nine years after Homestead. The workers who remained saw their daily wages shrink by one-fifth and their shifts extend to 12 hours. The samurai who funded Japan’s transformation lost their function, their stipends, and 40,000 of them their lives in a rebellion the new infrastructure crushed. China’s xiagang generation increased their precautionary savings by 40 percent, the arithmetic of a people who had learned that the state’s promises were conditional.
The molt succeeds for the institution. It does not succeed for the workforce. The institution absorbs the transition. The workforce absorbs the cost.This arrangement has held across the four cases examined here, spanning 250 years and three different systems of governance.
There is one question the historical record has never answered, because no one in a position to answer it has ever asked: is this an inevitability of institutional transformation or simply the price we have accepted, without deciding to accept it?
The 6 a.m. email arrived. The pattern is not new. Whether it keeps being absorbed by workers rather than borne by institutions is a choice the next institution will make, not a law it is obeying..
If you’d like to support the hours of research that go into these posts, you can support Precedent. Your contributions help keep this resource free and accessible for everyone.
PRECEDENT The satisfying explanation behind what just happened to you - traced across history, cultures, institutions, corporations.