Somewhere in the last minute, a chip in your phone woke up, did a few billion calculations, and went back to sleep. That chip was almost certainly built - layer by layer, atom by atom - on equipment made by a company most people have never heard of. Applied Materials does not design chips. It does not sell them at a store. It builds the machines that make the machines, and it has quietly done so since 1967.
If the semiconductor industry were a gold rush, Applied Materials would be the company selling the shovels, the sieves, the maps and the training - to every prospector on the field. When artificial intelligence sent chip demand soaring, Applied did not have to guess which chipmaker would win. It sells to all of them. That is the business, and in fiscal 2025 it produced a record $28.37 billion in revenue, a sixth consecutive year of growth.
01 / What it doesMaterials engineering, at the scale of atoms
A modern chip is not carved so much as it is grown. Hundreds of ultra-thin layers of metal and insulator are deposited onto a silicon wafer, then selectively removed, doped, polished and inspected - a recipe that can run to a thousand steps. Applied Materials makes the tools for most of those steps: deposition, etch, ion implantation, planarization, and metrology & inspection. Its machines can add or subtract material one atomic layer at a time.
The company calls this "materials engineering," and it is the part of chipmaking most outsiders overlook. Everyone talks about design and lithography. But turning a design into a working transistor is a physics problem measured in nanometers, and it is the problem Applied has spent nearly six decades solving.
02 / Who buys itIts customers make the chips in everything
Applied's customers are the world's leading logic, memory and foundry chipmakers, plus the manufacturers who build advanced displays. They operate the sprawling, dust-free factories called fabs across the United States, Taiwan, South Korea, China, Japan and Europe. When a foundry decides to build a new plant for a smaller process node, or a memory maker retools for the DRAM that feeds AI datacenters, they place orders with a short list of equipment companies. Applied is at the top of that list.
These are not casual purchases. A single leading-edge fab can cost tens of billions of dollars, and the equipment inside it is the largest line item. Once a chipmaker qualifies a tool for a given process, switching to a rival is slow and expensive, which is why the relationships tend to last for years and span multiple generations of technology. Applied's roughly 35,000 employees exist in large part to support that installed base - installing systems, training operators, and squeezing more yield out of tools that may run around the clock for a decade.
03 / The problem it solvesEvery nanometer smaller is a materials problem
For fifty years the industry kept its famous promise - more transistors, cheaper, every couple of years - largely by shrinking. But shrinking has gotten brutally hard. At today's dimensions, a few atoms out of place can ruin a chip, and simply making features smaller no longer delivers the gains it used to. The frontier has moved to new materials, new structures, and new ways to stack and connect chips.
That is precisely Applied's lane. Its tools enable the transition to next-generation transistors, higher-density memory, and advanced packaging that wires multiple chiplets together into one high-performance system. In a real sense, the company's job is to keep Moore's Law - or its modern successor - alive by engineering the materials no one else has figured out yet.
04 / How it's differentBreadth is the business model
The chip-equipment world is a tight club of five - Applied Materials, ASML, Lam Research, KLA and Tokyo Electron - that together control most of the market. Each has its stronghold. ASML owns lithography almost outright. KLA dominates inspection. Lam is a force in etch and deposition. What sets Applied apart is reach: it is the broadest of them all, with tools touching nearly every significant step of wafer fabrication.
That breadth is not an accident, it is the strategy. A customer building a new process can buy an interconnected suite from a single partner and let Applied co-engineer how the steps fit together. When the whole industry needs the same upgrade, breadth means Applied sells more of the upgrade than anyone.
05 / Products & servicesMachines, plus the services that keep them running
Applied's catalog runs from chemical and physical vapor deposition, atomic layer deposition and epitaxy, through etch, ion implantation and chemical mechanical planarization, to the e-beam and optical metrology tools that hunt for defects. Around the hardware sits Applied Global Services - spares, upgrades, remanufactured systems, and process-control and factory-automation software that keep an installed base of tools running at peak yield.
More recently the company opened the EPIC platform, a collaboration model where chipmakers, suppliers and researchers co-develop process technology, and it extended that idea to EPIC Advanced Packaging - the chiplets, through-silicon vias and interposers that AI systems increasingly depend on. In 2026 it introduced new systems aimed squarely at DRAM and advanced packaging for AI chips.
06 / Business modelSell the tool once, service it for a decade
The economics have two engines. The first is big-ticket capital equipment: fabs buy systems that can cost millions each. The second, and increasingly the prize, is recurring revenue. Every tool Applied has ever shipped needs parts, upgrades, and software - a growing, higher-margin service business that smooths out the notorious boom-and-bust of the equipment cycle. Revenue splits across three segments: Semiconductor Systems, Applied Global Services, and Display and Adjacent Markets.
07 / Expertise & cultureAn R&D company that happens to sell machines
Applied is, at its core, a research organization. Since Gary Dickerson took over as CEO in 2013, the company has poured a rising share of its operating budget into R&D - by some accounts from roughly 56% of operating expenses to about 69% - betting that whoever controls materials engineering controls the next decade of computing. The work is patient and deeply collaborative: engineers embed with customers for years to develop the recipe for a node that will not ship until later. The company also frames energy-efficient chipmaking as both a business and a sustainability goal.
08 / Where it fitsThe layer beneath the layer
Consumers see phones and laptops. Investors see NVIDIA and TSMC. Beneath all of it sits a thinner, quieter layer - the companies that make the manufacturing possible. Applied Materials owns more of that layer than anyone. It began in 1967 as a five-person Silicon Valley startup, four years before Intel shipped its first microprocessor, and grew into the world's largest semiconductor equipment company. It has no consumer brand to speak of. It just happens to be inside almost everything you use.
The founding story is pure early Silicon Valley. Michael McNeilly and a handful of co-founders started the company to serve a semiconductor industry that barely existed yet, opened an early office in Europe, and took the company public in 1972 on the Nasdaq under the ticker AMAT. For a stretch in the 1970s it wandered - even dabbling in unrelated businesses - before refocusing on the one thing it did better than anyone: building the equipment that makes chips. That focus carried it through the PC boom, the internet era, the smartphone decade, and now the age of AI.
The pattern across all of those cycles is the same. Each new wave of computing needs chips that are faster and more efficient than the last, and getting there is, underneath, a materials problem. Whoever solves that problem gets paid whether the winning device is a mainframe, a laptop, a phone or an AI accelerator. Applied has been on the paying side of that trade for nearly sixty years, which is the real reason a company almost no one can name keeps posting the best numbers in its history.