How Does the Semiconductor Supply Chain Work?

By Adam Patti, Chief Executive Officer, VistaShares | November 2026

The semiconductor supply chain is the sequence of specialized industries that turn a chip design into a working product. It runs from design software and intellectual property, through the equipment and materials used in fabrication plants, or fabs, to the manufacturing of logic and memory chips, and finally to packaging and testing. Each stage is concentrated among a limited number of companies and locations, which is why the industry is central to discussions of technology, trade and national security.

How large is the semiconductor industry?

The industry has grown rapidly with demand for AI computing. According to the Semiconductor Industry Association, global semiconductor sales reached $791.7 billion in 2025, up 25.6% from 2024, with logic products the biggest category by sales and memory second.¹ The SIA later reported revised 2025 sales of $795.6 billion, and cited a World Semiconductor Trade Statistics projection that sales could reach $1.5 trillion in 2026.² Forecasts may not be realized.

What happens at the design stage?

Chips begin as designs. Design involves defining a chip’s architecture, writing its logic and verifying that it works before any silicon is made.

Chip designers. Many companies design chips without manufacturing them. These are often called fabless companies. They create processors, graphics and AI accelerators, networking chips, and many other products.

Electronic design automation. Designers rely on specialized software to lay out billions of transistors and simulate how a chip behaves. A small number of companies supply most of this software.

Intellectual property. Many chips incorporate licensed building blocks, such as processor cores or interface designs, from intellectual property providers.

What equipment and materials does chipmaking require?

Fabrication depends on some of the most complex machines ever built.

Lithography. Lithography tools project circuit patterns onto silicon wafers. The most advanced chips require extreme ultraviolet lithography, a technology produced by a very small number of suppliers.

Deposition and etch. These tools add and remove thin layers of material to build up the structure of each chip, layer by layer.

Inspection and metrology. These systems measure and inspect wafers throughout production to detect defects.

Materials. Fabs consume silicon wafers, specialty gases, chemicals and photoresists, often supplied by specialized producers.

Equipment makers sell to every major chip manufacturer, so their revenue depends on the industry’s overall level of capital spending rather than on which chip designer succeeds.

How do fabs and foundries work?

A fab is a factory where chips are made. Building a fab for the most advanced chips costs many billions of dollars and takes years. Two business models exist:

Foundries manufacture chips designed by other companies. This model allows designers to avoid the cost of building fabs. A small number of foundries can produce chips at the most advanced process nodes.

Integrated device manufacturers both design and manufacture their own chips, and some also offer foundry services to others.

Process nodes describe successive generations of manufacturing technology. More advanced nodes allow more transistors in a given area, which can improve performance and energy efficiency. Each new node requires substantial investment in equipment and process development.

What role does memory play?

Memory chips store data. The two main types are DRAM, which provides fast working memory for processors, and NAND flash, which stores data in solid-state drives and devices. Memory is produced by a small number of large manufacturers and has historically been more cyclical than other parts of the industry, with prices rising and falling as supply and demand shift.

AI has increased demand for high-bandwidth memory, which stacks DRAM chips vertically and places them next to the processor so that data moves faster. High-bandwidth memory requires advanced manufacturing and packaging and is supplied by only a few producers.

What is advanced packaging?

After chips are made, they are cut from the wafer, tested and packaged so they can connect to a circuit board. Packaging has traditionally been a lower-value step. That has changed with AI.

Advanced packaging techniques place multiple pieces of silicon, such as a processor and several stacks of memory, close together in a single package with very dense connections. These designs allow components to work together as one large chip. Advanced packaging capacity has at times been a bottleneck for AI accelerators, and foundries and specialized packaging firms have expanded their capacity.

Where is the supply chain located?

The semiconductor supply chain is global but concentrated. Design and design software are concentrated in the United States. Leading-edge logic manufacturing is concentrated in Taiwan, and memory manufacturing in South Korea. Critical lithography equipment comes from Europe, other equipment from the United States and Japan, and many materials from Japan. Assembly, packaging and testing are concentrated in Asia.

U.S.-headquartered companies accounted for 53.4% of global semiconductor sales in 2025, the highest U.S. share since 1984, according to the SIA.² Governments in the United States, Europe and Asia have introduced incentives to expand domestic manufacturing, and export controls restrict sales of certain advanced chips and equipment.

What do semiconductor ETFs own?

Semiconductor ETFs typically hold companies across several stages: chip designers, foundries and integrated manufacturers, memory producers, and equipment makers. Some funds weight holdings by market capitalization, which concentrates assets in the biggest companies. Others cap individual positions or use equal weighting. Some include only U.S.-listed companies, while others include manufacturers and equipment makers in Asia and Europe.

When comparing semiconductor funds, investors can look at the mix across design, manufacturing, memory and equipment, concentration in the top holdings, geographic exposure and costs. Our article on what an AI ETF is describes how semiconductor exposure fits within broader AI funds.

What are the risks of the semiconductor industry?

Cyclicality. Chip demand and pricing have historically moved in cycles, with periods of shortage followed by oversupply.

Capital intensity. Fabs and equipment require very large investments, and returns depend on high utilization.

Concentration. Reliance on a few companies and locations creates operational and geopolitical exposure.

Trade and policy. Export controls, tariffs and subsidies can shift revenue and investment.

Technology risk. Each new process generation is harder and more expensive to develop.

Investments involve risk, including possible loss of principal. See our thematic ETF guide for related topics.

Frequently asked questions

What is the difference between a fabless company and a foundry?

A fabless company designs chips but does not manufacture them. A foundry manufactures chips designed by other companies. This division allows designers to focus on architecture and software while foundries invest in fabrication technology. Some companies, known as integrated device manufacturers, do both.

Why is lithography equipment so important?

Lithography defines the patterns that become a chip’s circuits. The most advanced chips require extreme ultraviolet lithography, which very few suppliers can produce. Without it, manufacturers cannot make chips at the most advanced process nodes, so the equipment is a critical point in the supply chain.

What is high-bandwidth memory?

High-bandwidth memory stacks DRAM chips vertically and places them beside a processor inside the same package, allowing data to move much faster than with conventional memory. It is essential for AI accelerators and is produced by only a few manufacturers using specialized techniques.

Why are semiconductors considered cyclical?

Chipmakers add capacity in large, expensive steps that take years to complete. When demand rises, shortages can drive prices up. When new capacity arrives or demand slows, oversupply can push prices down. Memory in particular has historically shown pronounced price cycles.

What does a semiconductor ETF typically hold?

Most semiconductor ETFs hold a mix of chip designers, manufacturers, memory producers and equipment makers. Weighting methods, concentration limits and geographic scope vary by fund, which can lead to meaningful differences between funds tracking the same industry. Reviewing the top holdings and the methodology clarifies what each fund owns.

This material is provided for educational purposes only. It is not investment, legal or tax advice, and it is not a recommendation or an offer to buy or sell any security. Information is drawn from sources believed to be reliable as of the date of publication, is not guaranteed as to accuracy or completeness, and may change without notice. Figures attributed to third parties are their estimates. Projections and forecasts are not guarantees of future results. Investors should consult their own financial, legal and tax professionals before making investment decisions. Investments involve risk, including possible loss of principal. Companies in a single industry or theme may be affected by common economic, regulatory or technological developments and may be more volatile than the broad market. References to companies, governments or programs are for illustration only.