How Intel Makes Chips: Why a Better Design Isn’t Enough

A processor fits in your hand. The system that makes it fills buildings. This documentary uses Intel’s own footage from Arizona, Oregon and New Mexico to follow that system and ask how Intel turns equipment, precision and coordination into products customers can depend on.

The footage comes from separate operations filmed at different times. It shows stages of manufacturing, not the journey of one traceable chip.

A wafer becomes a product

Chips start as wafers, thin discs of silicon. ASML, which makes lithography machines, describes chipmaking as a sequence of repeated steps: layers are added, patterns are exposed with light through a mask, also called a reticle, and material is removed where the design requires it. A chip does not come out of one machine in one pass. It comes out of a long run of operations that have to agree with each other.

The clean room suits, gloves and covered faces exist because tiny contamination can damage the structures being made. Wafers travel between tools in closed pods, which keeps that protection in place while they move.

The overhead network

Look up inside an Intel fab and you see its automated material handling system. Intel’s manufacturing press kit describes the one at its D1 factory in Oregon as more than 27 miles of multilevel track, with 1,700 autonomous vehicles carrying pods of up to 25 wafers through a fabrication journey that lasts weeks.

Speed alone does not solve the problem. If the next tool is not ready, an early arrival just waits. Different processing times make scheduling and temporary storage matter, which is why material handling suppliers such as Daifuku build storage into the system alongside transport. The goal is a useful arrival: the right material at the tool when production needs it.

Making good chips

Keeping equipment busy is not the goal either. The chips have to work. ASML describes how chipmakers measure printed patterns and inspect for defects, then use that information to improve accuracy and yield. Yield is the share of chips that work, and it connects the precision of the factory to the amount of useful output.

The packaging stage

Making the silicon is not the whole job. Chips also need packaging and connections to become part of a working system. In January 2024, Intel opened Fab 9 in Rio Rancho, New Mexico, for advanced packaging, including Foveros, its technology for stacking chips in three dimensions. Intel described Fab 9 as part of a previously announced $3.5 billion investment in its New Mexico operations. That figure covers the site investment, not the price of one machine.

What the factory sells

Each capability solves a different constraint. Transport feeds the tools, measurement stabilizes the process and packaging makes different product designs possible. Their value depends on how the whole operation performs, not on how impressive one machine looks on camera.

That is the logic behind Intel’s foundry ambition. A foundry manufactures chips for other companies, and a customer choosing one depends on far more than a promising design: process, packaging, capacity and support. At its Foundry Direct Connect event in April 2025, Intel’s chief executive, Lip-Bu Tan, said the company’s first job was to listen to customers and earn their trust. That trust is won through execution, which makes the factory part of what Intel sells.

It is the same lesson as Bosch’s Dresden wafer fab: the equipment is what you can see, but the value is in what the whole system delivers.

Which part of chip manufacturing surprised you most?