Silicon Valley is famous for creating the technology that powers our world. Companies based in this part of California design processors for computers, smartphones, artificial intelligence systems, and countless other devices we use every day. Yet there is something strange about the region that gave the world its name: many of the chips designed by Silicon Valley companies are manufactured somewhere else. Some of the world’s most advanced semiconductor factories are in Taiwan and South Korea, while new American factories are being built in places such as Arizona. Why would the home of so many chip designers send so much of its manufacturing work elsewhere?
The answer is more complicated than expensive land or high salaries. Semiconductor manufacturing has changed so much that designing a chip and producing it have become two very different businesses. Silicon Valley became exceptionally good at designing chips, developing technology, and building companies around new ideas, while other regions invested heavily in the factories needed to turn those ideas into physical products. Over time, the economics of the industry encouraged companies to keep their engineers in California and manufacture their chips elsewhere.
Silicon Valley Was Once a Place Where Chips Were Made
The story makes more sense when we remember what Silicon Valley looked like in its early years. During the 1950s and 1960s, semiconductor manufacturing was one of the main reasons technology companies came to the region. Fairchild Semiconductor, founded in 1957, became a major force in the development of integrated circuits and helped establish a culture in which engineers could leave established companies and start businesses of their own. Intel followed in 1968, founded by former Fairchild employees Robert Noyce and Gordon Moore. These companies were involved in both developing new technology and figuring out how to manufacture it reliably. The region’s reputation grew because ideas could move from engineering teams into real products.
The connection between research and manufacturing mattered enormously. When a new chip failed during production, engineers needed to understand why it had failed and how to correct the problem. When a manufacturing process improved, designers could use that knowledge to develop better products. Having these activities close together made it easier to exchange information and solve technical problems. Manufacturing was not merely the final step after invention; it helped engineers discover what was possible in the first place.
You can still see the results of that history in Silicon Valley today. The region remains home to semiconductor designers, equipment companies, research teams, and people with decades of technical experience. However, an industrial region does not keep the same role forever just because it was successful in the past. As chip manufacturing became more expensive and specialised, companies began looking for different ways to organise their businesses. The industry gradually separated activities that had once been performed under the same roof.
The Real Turning Point Was the Cost of Making Chips
A modern semiconductor factory, usually called a fabrication plant or fab, is one of the most demanding industrial facilities to build. It requires specialised cleanrooms, advanced machinery, carefully controlled air conditions, water purification systems, and dependable electricity. The equipment used to manufacture advanced chips can cost tens of millions of dollars per machine, while an entire leading-edge fab can require an investment of tens of billions. A company making that investment must be confident that it can produce enough chips, sell them at competitive prices, and keep the facility useful as technology advances.
The problem becomes clearer when you look at how quickly manufacturing technology changes. A factory designed for one generation of chips may not be suitable for the next generation without major upgrades or new equipment. Smaller circuit features require more advanced production techniques, and improvements often involve expensive machinery and difficult engineering work. Companies cannot simply build a factory once and assume that it will remain competitive for the next thirty years. They must continue investing in equipment, process development, and production improvements.
This creates a difficult choice for companies that design chips. Should they spend billions of dollars building and maintaining their own factories, or should they concentrate on designing better products and pay another company to manufacture them? For many businesses, especially smaller ones, the second option makes more financial sense. They can direct more money towards research, software, engineering talent, and product development instead of trying to compete in manufacturing at the same time. That decision helped change the structure of the semiconductor industry.
How Silicon Valley Learned to Design Chips Without Making Them
One of the biggest changes in the industry was the growth of the fabless business model. A fabless company designs semiconductor products but does not operate its own high-volume wafer fabrication plants. Instead, it sends its designs to a specialised manufacturer known as a foundry, which produces chips for other companies. This arrangement allows businesses to develop sophisticated processors without paying the enormous cost of building their own advanced manufacturing facilities. It also allows foundries to serve several customers using the same expensive infrastructure.
Taiwan Semiconductor Manufacturing Company, better known as TSMC, helped make this model successful after its founding in 1987. Instead of competing directly with its customers by selling its own branded chips, TSMC focused on manufacturing chips designed by other businesses. This gave chip designers a specialised manufacturing partner and allowed the foundry to spread its investment across many customers. As demand grew, TSMC could invest in larger factories and more advanced production technologies.
The model was especially useful for Silicon Valley startups. A small company could hire engineers, develop a new chip, and seek customers without first raising enough money to build a factory. That opened the door to more experimentation and allowed new businesses to compete in areas such as networking, mobile computing, and artificial intelligence. The companies could focus on what they did best while relying on specialist manufacturers for production. In a surprising way, the growth of overseas manufacturing helped Silicon Valley create more semiconductor companies.
This is an important part of the answer that often gets missed. Silicon Valley did not simply abandon manufacturing because it became too expensive. The rise of specialised foundries gave companies a practical alternative, and many chose to use it. Once that business model became established, there was less reason for every chip designer to maintain a factory of its own. The region’s success in chip design could continue even as more physical production took place elsewhere.
Why Taiwan Became So Important
Taiwan’s position in semiconductor manufacturing was built over decades. Companies and public institutions invested in technical skills, industrial infrastructure, manufacturing capacity, and the development of local suppliers. TSMC became particularly important because it focused on producing chips for other companies and developed the scale needed to serve customers around the world. Its factories could manufacture products for businesses that competed in different markets, from consumer electronics to high-performance computing.
Scale matters because a semiconductor factory has enormous fixed costs. The company must pay for its buildings, equipment, engineering teams, and production systems whether the factory is operating at full capacity or not. Producing more chips across a wider customer base can help spread those costs across a larger number of products. A foundry serving many customers may therefore have an advantage over a company that builds a factory mainly to produce its own chips. The advantage becomes stronger when the foundry also develops years of experience in improving manufacturing yields and solving production problems.
Manufacturing experience is easy to underestimate. A factory may have advanced equipment, but that does not mean it will immediately produce chips efficiently. Engineers must learn how to control defects, improve the number of usable chips produced from each wafer, and keep the process stable at high volume. Small improvements can have a large financial effect when a factory produces millions of chips. An established manufacturer has accumulated knowledge that a new competitor cannot buy simply by purchasing the same machines.
Taiwan also developed a network of businesses and workers that support semiconductor production. Equipment suppliers, chemical companies, specialist engineers, and other service providers can work closely with manufacturers. This makes it easier to maintain production and solve technical problems quickly. Once a region has a strong manufacturing ecosystem, it becomes more attractive to companies considering their next factory. New investment strengthens the ecosystem, which can then attract still more investment.
Why Silicon Valley Is an Expensive Place to Build a Fab
The cost of manufacturing helps explain the industry’s global shift, but the location of the factory matters too. Silicon Valley has high property prices, expensive housing, and strong competition for skilled workers. A company planning a large fab needs a suitable industrial site with enough space for the building, utilities, chemical systems, water treatment, and possible future expansion. Finding such a site in a heavily developed region can be difficult and costly.
The financial calculation does not end when the company buys the land. A fab needs reliable supplies of electricity and water, along with transport connections and access to skilled employees. Water must be treated to extremely high purity before it can be used to clean silicon wafers, and electricity must remain dependable because interruptions can disrupt production. The company must also consider maintenance, replacement equipment, waste treatment, and other operating expenses. These costs continue throughout the life of the facility.
However, it would be misleading to blame Silicon Valley’s high wages for the entire problem. Research reviewed by the National Academies of Sciences, Engineering, and Medicine indicates that differences in labour costs are not the main explanation for the historical gap in semiconductor fabrication costs between the United States and some Asian locations. Government subsidies and the treatment of large factory investments have played an important role. This matters because a company comparing two possible sites is looking at the total cost of building and operating a factory, including financial support and tax treatment. A location with higher wages may still be competitive if other costs are lower or public incentives are stronger.
Silicon Valley also competes with industries that do not need enormous factories. A software company can often expand by hiring engineers and leasing office space, while a leading-edge fab needs specialised industrial infrastructure and a huge amount of capital. Technology companies in the region can earn substantial returns from software, chip design, and intellectual property without taking on the financial risks of manufacturing. That makes it harder for a new fab to compete for investment against other possible uses of the same money.
Land, Water, Electricity, and Permits All Matter
Large semiconductor factories require more than an affordable building site. They need reliable utilities and a clear path through the approvals required for construction and operation. California’s environmental review process can require detailed studies of a project’s effects on its surroundings, depending on the site and the approvals involved. These reviews can protect communities and identify problems before construction begins, although uncertainty and delays can also increase project costs.
Water is an especially important issue because semiconductor manufacturing uses highly purified water at several stages of production. Factories need treatment systems that remove particles and unwanted substances, along with systems for managing wastewater and recovering water where possible. California has experienced repeated droughts, so a manufacturer must examine the long-term reliability of water supplies before committing to a major facility. Arizona faces serious water challenges too, which means the decision cannot be reduced to the idea that one state has water and the other does not.
Electricity is another basic requirement. A fab needs continuous power for manufacturing tools, cleanroom ventilation, cooling, water treatment, and monitoring systems. Companies must consider not only the price of electricity but also the reliability of the local grid and the time needed to provide enough capacity for a new industrial campus. A location that cannot support the factory’s power requirements on schedule may lose the investment to a competing region.
These issues do not make semiconductor manufacturing impossible in California. They mean that a proposed project must fit the region’s available land, utility infrastructure, environmental requirements, and workforce. The best site is not necessarily the cheapest piece of land or the place with the lowest electricity bill. It is the location where the full combination of requirements can be met at a cost and level of risk the company is willing to accept.
Why Arizona and Texas Are Attracting New Factories
Arizona has become an important destination for semiconductor investment because it offers large industrial sites and has developed a growing manufacturing workforce. Intel already has a major manufacturing presence in the Phoenix area, and TSMC selected the region for a substantial investment in advanced chip production. The US Department of Commerce announced up to $6.6 billion in direct funding for TSMC Arizona under the CHIPS and Science Act, supporting a planned investment of more than $65 billion in three fabrication plants. The project shows how federal incentives can influence where companies decide to build.
The attraction is not simply that Arizona has more available land than Silicon Valley. A semiconductor project needs trained workers, reliable utilities, transport infrastructure, suppliers, and long-term support from local institutions. Building these capabilities takes time, and the region must continue developing them as new factories come online. A large investment can help create a manufacturing ecosystem, but construction alone does not guarantee that the ecosystem will succeed.
Texas offers another example of how a region can develop semiconductor capabilities over time. Austin has attracted semiconductor companies and built links between industry, universities, research organisations, and the local workforce. These connections help companies find technical talent and give workers opportunities to gain experience in manufacturing. The presence of established businesses can also encourage suppliers to locate nearby, making the region more attractive to future investors.
None of these locations is free from difficulties. Arizona must manage water availability, and every large industrial project needs to consider its effects on local infrastructure and communities. The point is that semiconductor companies compare complete locations, not just individual costs. A region that can offer land, utilities, workers, suppliers, and public support in a coordinated way has a stronger chance of winning a major factory investment.
The CHIPS Act Is Trying to Change the Economics
The United States has become concerned about relying too heavily on overseas manufacturing for semiconductors. Chips are essential to cars, communication systems, medical equipment, data centres, industrial machinery, and military technology. When supplies become limited, the effects can spread across many industries. The shortages experienced during the COVID-19 pandemic made these risks more visible to businesses and policymakers.
The CHIPS and Science Act, signed into law in 2022, was designed to encourage more semiconductor manufacturing and research in the United States. It provided $39 billion for manufacturing incentives, alongside funding for research and other semiconductor programmes. The goal was to encourage companies to build or expand facilities, strengthen domestic capabilities, and reduce vulnerabilities in the supply chain. Government support can make a project more attractive by reducing some of the financial disadvantages of building in the United States.
The results will take years to measure properly. Announcing a factory is not the same as completing it, installing equipment, training workers, and producing chips at a competitive cost. A factory must also keep investing as manufacturing technology advances. Public funding can help companies make the initial commitment, but it cannot guarantee that every project will succeed or that domestic production will immediately match the cost and experience of established overseas manufacturers.
For Silicon Valley, the legislation creates opportunities but does not settle the question of where factories should be built. Some companies may find California suitable for particular manufacturing projects, while others may prefer Arizona, Texas, Oregon, or another state. Federal incentives can change the calculation, but land, utilities, workforce availability, and the surrounding industrial ecosystem still matter. The goal is not simply to build more factories anywhere; it is to create manufacturing capacity that can operate successfully over the long term.
Silicon Valley Still Matters to the Semiconductor Industry
It would be a mistake to assume that Silicon Valley has become unimportant to chipmaking. The region remains a major centre for semiconductor design, engineering, research, and the development of manufacturing equipment. These activities are essential because factories cannot produce a new generation of chips without designs, software tools, process knowledge, and specialised machinery. Many of the companies that support semiconductor production continue to have important operations in the Bay Area.
Intel is one example of the region’s continuing connection to manufacturing. Its Santa Clara operations include headquarters activities and Mask Operations, which produces photolithography masks used to transfer circuit patterns during chip manufacturing. These masks are part of the process that turns a design into a physical circuit on a silicon wafer. Such work shows that semiconductor manufacturing involves many specialised activities beyond running a large wafer fab.
The distinction is important because people often use the word manufacturing to describe only the factory where chips are made on silicon wafers. The wider industry also includes equipment production, materials, chip design, packaging, testing, and research into new processes. Silicon Valley can remain highly influential in these areas even if a smaller share of the world’s chips is physically fabricated nearby. Its engineers and companies contribute to manufacturing systems used in factories across the world.
There is also a practical reason to keep research and manufacturing connected without requiring them to share the same location. Design teams can work closely with foundries through technical collaboration, testing, and process development. A chip designed in California can be manufactured in Taiwan or Arizona while engineers in both locations work to improve its performance and reliability. Distance creates challenges, but modern communication and specialised partnerships make this arrangement possible.
Could Semiconductor Manufacturing Return to Silicon Valley?
Some manufacturing could expand in Silicon Valley, but a major return of large leading-edge fabs would require more than a desire to restore the region’s industrial past. Companies would need suitable sites, dependable utilities, trained workers, predictable project timelines, and a strong business case for investing billions of dollars. California would also need to consider how large industrial projects fit into its housing, transport, water, and environmental plans. These are practical questions that must be answered before a company commits to a factory.
The region may have stronger opportunities in specialised manufacturing, pilot production, semiconductor equipment, and advanced packaging. Pilot lines allow engineers to test manufacturing processes before committing to full-scale production, while advanced packaging brings multiple chip components together into a single system. These activities can benefit from Silicon Valley’s engineering talent and research environment. They will not replace every capability of a leading-edge fab, but they could strengthen the region’s contribution to the industry.
Universities and companies could also work more closely to train manufacturing engineers and technicians. New factories need workers who understand production equipment, process control, maintenance, safety, and quality assurance. Training programmes can help develop these skills, but they must be connected to real employment opportunities and the needs of manufacturers. A sustainable industrial strategy requires people and institutions, not just investment announcements.
Ultimately, Silicon Valley does not need to reproduce every part of the semiconductor supply chain to remain important. It needs to understand where its strengths offer a genuine advantage and where other regions are better equipped to handle large-scale production. If California can make industrial investment more predictable and improve access to suitable sites and infrastructure, some manufacturing projects may become more attractive. The decision will still depend on whether the long-term benefits justify the costs.
The Bigger Lesson Behind Silicon Valley’s Manufacturing Gap
The story of semiconductor manufacturing in Silicon Valley is a reminder that technological leadership and manufacturing leadership are not the same thing. A region can develop outstanding products while another region becomes better at producing them at scale. Over time, companies organise their businesses around those differences, especially when manufacturing requires enormous investment and specialised experience.
Silicon Valley’s rise as a centre for chip design was partly enabled by the growth of foundries elsewhere. That model allowed startups to develop new products without owning factories, while specialised manufacturers built expertise and spread their costs across many customers. Government incentives and industrial policies helped shape where the factories were built, and established supplier networks made some locations increasingly attractive. The change was the result of business decisions and public investment working together over decades.
Bringing more semiconductor manufacturing to the United States will therefore require more than constructing buildings. It will require competitive financing, reliable infrastructure, skilled workers, experienced suppliers, and the ability to improve production over time. New factories in Arizona and other states may help rebuild domestic capacity, but they will need years of investment and learning to reach their full potential. The same is true for any future manufacturing project in California.
So why isn’t there more semiconductor manufacturing in Silicon Valley? Because the region became exceptionally good at designing chips while the economics of fabrication encouraged production to concentrate in specialised manufacturing centres. High local costs and the difficulty of expanding industrial infrastructure make large factories harder to justify, but they are only part of the explanation. The deeper reason is that the semiconductor industry developed a global division of work, and Silicon Valley adapted to it. Whether more manufacturing returns will depend on whether California can offer the conditions that modern chip production requires.
Leave a Reply