The few companies that control the world's chip supply

The semiconductor industry is often described as competitive, and at the level of finished products it is. At the level of the supply chain that produces those products it is one of the most concentrated industries on earth. A single company makes the machines without which the most advanced chips cannot be manufactured. A single company makes most of those advanced chips for everyone else. A single company designs the processors that the artificial intelligence boom runs on. Strip the market back to its chokepoints and you are left with a short list of firms that, between them, hold the modern economy in their hands.

What follows is a map of those chokepoints, layer by layer, because the concentration is not uniform. It sits at very specific points in the chain, and those points are where the real power and the real risk both live.

The manufacturing chokepoint

Taiwan Semiconductor Manufacturing Company makes the majority of the world's chips by value and the overwhelming majority of the most advanced ones. When a designer wants a leading edge processor built, there is effectively one supplier with the yield, the scale and the process maturity to do it at volume. Samsung is the only credible second source at the leading edge, and it sits well behind on market share. This single point of concentration is why Taiwan has become the most strategically sensitive location in global trade, and why every major government is now spending heavily to build domestic capacity it does not currently have.

The equipment chokepoint

Upstream of the foundries sits an even narrower bottleneck. ASML of the Netherlands is the only company in the world that makes extreme ultraviolet lithography machines, the equipment required to print the smallest features on a modern chip. There is no second supplier. A foundry that wants to build at the leading edge must buy from ASML, which means a single European firm holds a veto over the entire advanced manufacturing layer. Alongside it, Applied Materials, Lam Research and Tokyo Electron dominate the other critical steps of deposition and etching. The tools layer is arguably more concentrated than the chips layer itself.

The design chokepoint

At the design layer the concentration shifts by category. Nvidia designs the graphics and accelerator chips that the entire artificial intelligence build out depends on, and holds a commanding share of that market. Its position is reinforced by software rather than silicon alone, because the developer ecosystem around its platform raises the cost of switching to a rival. In central processors the long standing pair of Intel and AMD still set the terms for personal computers and servers on the dominant instruction set, while Arm licenses the architecture that sits inside almost every mobile device and a growing share of everything else. Few companies design the brains of modern computing, and each owns a category that is hard to contest.

• TSMC manufactures the majority of the world's chips by value and most leading edge production.
• ASML is the sole maker of the extreme ultraviolet lithography machines needed for advanced chips.
• Nvidia holds a commanding share of the accelerators that power artificial intelligence.
• Samsung is the only credible second source at the leading edge of manufacturing.
• Intel, AMD and Arm between them define the processor architectures of modern computing.
• Applied Materials, Lam Research and Tokyo Electron dominate the other critical equipment steps.

The pattern across all three layers is the same. Competition exists in the abstract, but at each specific chokepoint there is one clear leader and at most one viable alternative. That is the defining feature of the industry, and it is why a disruption at any single node propagates through the entire global economy within weeks.

Why the concentration is so durable

These positions are not held by accident, and they are not easily contested. A leading edge fabrication plant now costs more than twenty billion dollars and takes years to bring to full yield. The lithography machines inside it are among the most complex objects ever manufactured, built from supply chains that themselves took decades to assemble. The design leaders are protected by software ecosystems and patent libraries that a new entrant cannot replicate quickly. Each layer therefore has a moat made of capital, time and accumulated know how, and all three are in short supply. That is why government money aimed at building alternatives produces results slowly, if at all.

Why it matters beyond the sector

Chips are no longer just an input to electronics. They are the input to artificial intelligence, to advanced weapons, to electric vehicles, to energy infrastructure and to the data centres that run the global economy. Control of the chokepoints described above is therefore control of the commanding heights of the next decade of growth. This is the reason semiconductors have moved from a business story to a national security story, and why export controls, subsidies and investment screening now shape the sector as much as ordinary commercial demand does.

Our reading

We see three durable conclusions in this map. First, the chokepoints in equipment and leading edge manufacturing are the hardest to dislodge and therefore the most strategically valuable, which is where pricing power and political leverage concentrate. Second, the design layer is more contestable than it looks, because software moats can erode as customers actively fund alternatives to avoid dependence, so the leaders there must keep earning their position. Third, the wave of government spending aimed at diversifying supply will broaden the map over time but will not break the core concentration this decade, because capital and time cannot be compressed past a point. For a long horizon investor the sector is best understood not as a single trade but as a chain of distinct chokepoints, each with its own moat, its own risk and its own exposure to the politics now reshaping the industry.