Who gets paid first when a company announces a multibillion-dollar chip factory? Not the chipmaker. The first checks go to the companies that build the machines inside the building, because a fab is mostly a shell until the tools arrive.
That is why semiconductor equipment makers have become one of the clearest ways for investors to trade the global scramble for new chip capacity. Governments in the United States, Europe, Japan, South Korea and elsewhere are subsidizing domestic production, while the demand for artificial intelligence hardware is pushing chipmakers to expand at the leading edge. Each new plant needs lithography systems, deposition and etch tools, inspection gear and cleaning equipment before it can produce a single wafer.

Why the Toolmakers Get Paid First
A chip factory is built in a predictable sequence. The construction crews pour concrete and install cleanrooms, then the equipment is ordered, shipped and calibrated over many months. Equipment makers book revenue well before the fab produces anything, which means their order books can swell long before the chipmaker’s own earnings show any benefit.
The supplier landscape is also unusually concentrated. Extreme ultraviolet lithography, the technology needed to print the finest features on advanced chips, is supplied by a single company, ASML. In other categories a small group of firms, including Applied Materials, Lam Research, Tokyo Electron and KLA, dominate deposition, etch and process control. When a market has few suppliers and very high technical barriers, customers have little room to negotiate, and the suppliers tend to keep healthy margins even when demand softens.
Complexity adds to the pricing power. Advanced chips require hundreds of process steps, and each new generation of transistor design tends to need more of them, not fewer. Features are stacked, etched deeper and measured more often. A fab that moves to a more advanced node does not just replace old tools, it typically buys more tool time per wafer. That is the structural reason equipment spending can grow faster than chip output.
The Installed Base Keeps Paying
The sale of a machine is only the opening transaction. Every tool placed in a fab becomes a long-term customer relationship, with spare parts, upgrades, software and maintenance contracts that continue for as long as the machine runs. This service revenue is steadier than new equipment sales, which makes a buildout wave more valuable than the headline order numbers suggest.
It also gives toolmakers a cushion that pure hardware businesses lack. When a chipmaker delays a new fab because memory prices have slumped, the machines already running in older plants still need to be serviced and upgraded. The more tools the industry installs during a buildout, the larger that recurring base becomes for the next downturn.

Where the Cycle Can Bite
None of this makes the sector immune to the boom-and-bust rhythm that has defined semiconductors for decades. Chipmakers order tools when they are confident about demand and cancel or delay when they are not. Memory makers in particular have a history of overbuilding, watching prices collapse, and then slashing capital spending almost overnight. Equipment suppliers sit directly downstream of those decisions, so their revenue can swing harder than the chip industry’s own.
Geopolitics adds a second layer of risk. Export controls on advanced tools sold to Chinese customers have reshaped what toolmakers can ship and to whom, and China has been a large buyer for several of these companies. Some of that demand was pulled forward as buyers stocked up ahead of tighter rules, which can flatter recent results and leave a gap later. Any new restriction, or any retaliation that targets the supply chain, lands on the income statements of the equipment firms first.

Then there is the question of how much of the buildout is real demand and how much is subsidy-driven optimism. Government incentives can make a fab economically viable in a location where it otherwise would not be, but they cannot guarantee that the plant will find enough customers once it opens. Some announced projects have already slipped or been scaled back, and the tool orders attached to them move with the schedule. A stock that prices in every announced factory is assuming that all of them get built on time, and history suggests that rarely happens.
The sharper test comes when the first wave of new plants is running and the question shifts from how many tools to buy to how many chips to sell. If AI demand keeps absorbing advanced logic and high-bandwidth memory, the orders keep coming. If it stalls while new capacity is still arriving, the toolmakers will be the first to feel the pullback, and their valuations already assume it will not come.






