The Chain of Price-Squeezing: Why Korea Never Had a Narrative of Expansion

When you study history, what is absent often says more than what is present.

Korean history has no narrative of expansion. More precisely, it is scarce. Most of it is a story of defense — enduring invasion, holding out, recovering. That King Gwanggaeto is summoned so constantly is itself proof of the absence: if memories of expansion were common, no single one would be called up so repeatedly.

Read this as lament and you stop there. I read it differently. The absence of an expansion narrative is not a question of national temperament but of structure. And structure can be analyzed. In this essay I want to descend through that structure one level at a time — starting from the chaebol, passing through the funding structure of state power, down to the very end of the price sheet. And from that end, I will look back up. There hangs a larger chain in which even Samsung is a subcontractor.

First, let us set down a lens.

The way to hold the board in world history has been swapped out era by era. In antiquity it was land: Rome bound the Mediterranean with law and roads. In the medieval world it was institutions and religion. In the early modern era it was trade and navies — the Netherlands was a small country, but it held the board with the invention of the joint-stock company. In the industrial age it was factories and capital. In the twentieth century it was standards, the reserve currency, and technical protocols. America wove the world together with the dollar and internet standards.

Land → trade → capital → standards → protocols. The medium of expansion kept moving.

This movement matters for one reason: when the medium changes, the hierarchy flips. The player who was strong when land was the medium gets pushed aside when standards become the medium. Conversely, a small player disadvantaged on land can stand at the top in finance or standards. The Netherlands did.

Most of the history in which Korea “failed to seize the initiative” belongs to the era when the board’s medium was land and military force. Small, wedged in, always between a larger continental power and a maritime power. In a game where territory is the medium, Korea’s geography was a structural disadvantage. That was a condition that could not be changed.

But the medium has already moved on. What holds the board today is not land but IP, standards, and chokepoints. With that premise laid down, let us descend into Korea’s internal structure.

Seen from a distance, Korea’s power structure shows two axes.

One axis holds the narrative. The government binds people with the language of patriotism: “we must unite to survive,” “the national interest,” “security.” The other axis holds the capital. The actual cash and technology sit with a handful of groups like Samsung and SK Hynix.

These two do not compete. They need each other. The government cannot produce results — exports, employment, GDP — without the chaebol, and the chaebol cannot obtain regulation, policy, and absolution without the government. So they hold each other up.

Here I use the word “cartel” with care. This is not backroom collusion. Frame it that way and you collapse at the first “show me the evidence.” What I mean is something sturdier — a state in which each side merely followed its own incentives, and those paths converged on an equilibrium where each reinforces the other. The structure holds even if no one is evil. In fact, it is more stable precisely because there are no villains.

And the function the patriotic narrative serves in this structure is the crux. The patriotic narrative operates as a substitute for the expansion narrative. The defensive language of “we must protect” legitimizes preserving the current structure. “Let us protect what we have built” means “let us not shake the board” — which is exactly what this equilibrium wants. The language of expansion, by contrast — “let us set a new board, let us go outward” — is dangerous. It shakes the existing channels.

So the defensive narrative reproduces itself and the expansion narrative is suppressed. The “absence of an expansion narrative” from chapter one may be no accident. The narrative-capital equilibrium keeps generating the language of defense.

To see how this structure works, let us put one node under the microscope: the National Intelligence Service.

People usually think of the NIS simply as a powerful agency. By budget it is indeed not small. Its main budget for 2024 alone runs to roughly 900 billion won, and once you add the intelligence budgets tucked inside other agencies, the total is estimated to exceed one trillion — no one knows the exact figure. The entire budget is classified as special activity funds, structured so that outside oversight is practically impossible.

But being large and being self-sufficient are different things. The NIS cannot earn its own money. The chaebol have their own cash flows; the NIS’s funding is entirely subordinate to allocation. It depends one hundred percent on a budget the National Assembly can cut every year.

Here a paradox arises. The very absence of self-generated funding becomes a structural incentive to attach itself to outside resources. A self-sufficient agency has no reason to attach. A dependent agency must. And the real resources of modern intelligence work — semiconductors, telecommunications, data, technology — sit with private conglomerates.

Three things overlap: opacity (no one outside can see in), funding dependence (it cannot earn for itself), and strong authority (it has things to give — intelligence, investigation, licensing). This triad is the ideal condition for capital and authority to be traded quietly. Authority without money and money without authority fill each other precisely.

Let me be clear. This is not the factual claim that “a cartel exists.” It is the structural diagnosis that “this kind of structure easily breeds cartels.” I do not cross that line. Structural analysis needs no evidence; factual claims do. The NIS is just one scene showing how the narrative-capital division of labor attaches at the micro level.

Now let us see who pays the cost.

When narrative and capital fix the board from above, the cost of that stability is passed downward. Large firms, as primes, cut the unit prices paid to their small and mid-sized subcontractors. The subcontractors cannot leave. There is no other channel to market — this is the consequence of the blocked expansion channel. Leaving means dying, so they stay, even while being squeezed.

So margins are sucked upward, and the subcontractors are left with no capacity for R&D or wage increases. Here is the decisive point: price-squeezing is not mere unfairness — it is a device that physically seals off the expansion channel. Without margin, there is no seed money to invest in your own technology, brand, or overseas channels. Without seed money, you cannot step outside the prime. Unable to leave, your dependence deepens. This cycle severs subcontractors from any possibility of becoming players who set the board.

This is how the burial of the expansion narrative is executed at the level of individual firms. The narrative does not vanish on its own; there is a concrete hand that buries it. That hand is the price sheet.

Here too, let us keep our balance. Painting this as “evil big firms exploiting virtuous small ones” is half a picture. The primes are also moved by the logic of survival. Squeezed on costs by global competition, they pass it downward. The primes are being squeezed from above as well. So the accurate picture is this — there is a chain along which cost pressure flows from top to bottom, and at its very end there is a layer with nowhere left to pass it on, a layer consumed not by capital accumulation but by survival. Not villains but structure flushes the cost to the bottom.

But the chain does not end at the border.

The large firms that are primes at home are subcontractors in the global chain. When Nvidia holds GPU design, supply, and priority, Samsung and Hynix too become subcontractors beneath it, conforming to specs and having their volumes adjusted. Delivering HBM, falling behind in the foundry race. The chain continues:

Nvidia → Samsung and Hynix → Korean mid-sized and small firms → those below them.

Each layer is squeezed from above and passes it down. The domestic structure was in fact one segment of a larger chain.

And here is the decisive question. Why is Nvidia at the top? Not because it manufactures best. Nvidia has no factories. It is fabless. No land, no plants — yet it sits at the top of the chain. What does it hold?

It holds the chokepoint. Design architecture, a software standard (CUDA), ecosystem lock-in. In other words, exactly what the lens showed — the board’s medium has moved from manufacturing to standards and IP. Samsung holds manufacturing, the previous era’s medium; Nvidia holds the standard, this era’s medium. So no matter how well you manufacture, you slot in beneath whoever holds the standard.

The very fact that the manufacturing champion sits beneath the standards champion is proof that the medium has already passed over. The lens is playing out, as is, in the semiconductor chain.

Now we return to the opening question. In a structure whose expansion channels are sealed, what is the way through?

The answer is not to go deeper down inside the chain. Contract manufacturing, API reselling — take those seats and you are squeezed forever. You are simply consumed by survival at the end of the line, with nowhere further to pass the cost.

The only way through is to hold a chokepoint. Just as Nvidia stood at the top without factories, holding standards, IP, and control lets you stand upstream in the chain even at a small scale. It is the only asymmetric strategy in an era whose medium has moved to IP. It is the modern version of how the Netherlands held the board without land.

This is why I am building an on-device inference engine. What regulated industries, the public sector, and defense want is AI they control, where their data never leaves. The further a country sits from the frontier, the weaker an institution’s hand, the greater the demand for this sovereignty. Control freed from cloud dependence — that is how even small institutions and individuals come to hold a piece of the chokepoint.

Bind the IP with patents, hold the grounds of performance with a runtime of your own, productize control through on-prem. This is the opposite direction from contract manufacturing. It is an attempt to stand not in the layer that gets squeezed, but in the layer that cannot be squeezed.

Korean history’s lack of an expansion narrative is a story from the era when the medium of expansion was land. The medium has already passed to IP, standards, and control. For a player pushed off the board of land, this is the only window that has ever opened.

We began from what is absent. The absent narrative is now ours to write.