From a Taipei Breakfast to a Global Chip Powerhouse: How Taiwan Built Its Semiconductor Industry

A small breakfast bill, a NT$400 million decision

Seven officials and engineers gathered at a modest Taipei breakfast shop one morning in February 1974. Their soy milk and fried dough cost less than NT$400. The plan they agreed on called for NT$400 million to bring integrated circuit technology to Taiwan, train engineers overseas and build a demonstration factory.

That decision became the RCA Project, the starting point for a manufacturing base that grew into United Microelectronics Corporation, known as UMC, and later Taiwan Semiconductor Manufacturing Company, or TSMC. Fifty years later, chips manufactured in Taiwan power smartphones, data centers and artificial intelligence systems around the world.

Taiwan's transformation began with a practical response to economic and political pressure. The oil shock of 1973 had raised costs and strained an economy built around exports. Diplomatic setbacks had also left the island increasingly isolated in world politics. Premier Chiang Ching-kuo asked his aides to find a technology that could give Taiwan a stronger basis for growth.

Pan Wen-yuan, a Chinese American engineer who led work at RCA Laboratories in the United States, proposed moving from assembly work that depended heavily on labor into integrated circuits. These are chips that bring electronic components together on a small piece of semiconductor material. Economics minister Sun Yun-hsuan supported the proposal, while Executive Yuan secretary general Fei Hua helped drive it forward.

The ambition went beyond assembling or packaging chips made elsewhere. Taiwan would learn every stage of manufacturing, develop its own engineers and improve the technology after bringing it home. The breakfast decision joined political backing to a concrete industrial task: build the capability to make chips, then turn that capability into businesses.

Who turned the idea into a national project?

The seven participants brought together expertise in government, telecommunications, research and industrial technology. Their roles connected the proposed chip program to institutions that could fund it, organize training and put it into production.

• Sun Yun-hsuan, minister of economic affairs.
• Kao Yu-shu, director general of Transportation and Communications.
• Wang Chao-chen, president of the Industrial Technology Research Institute.
• Kang Pao-huang, director of Telecommunications Laboratories.
• Fei Hua, secretary general of the Executive Yuan.
• Fang Hsien-chi, director general of Telecommunications.
• Pan Wen-yuan, RCA Laboratories director.

Pan spent ten days shut away at the Grand Hotel writing the plan. It received approval in August 1974 despite criticism that semiconductor manufacturing required too much capital and was too complicated for Taiwan to undertake. Sun pressed ahead with the program.

The government assigned the Industrial Technology Research Institute, or ITRI, to oversee the work through its Electronics Research and Service Organization, known as ERSO. A Technology Advisory Council brought together experts based overseas to guide technology choices and keep the project connected to market demand.

The council focused on consumer electronics, where Taiwan's manufacturers could find customers quickly. Its chosen technology was complementary metal oxide semiconductor, or CMOS, valued for low power consumption and reliability. The program had a clear sequence: acquire complete manufacturing capability, learn to operate it, improve it and spread that knowledge through new companies.

A technology transfer contract with RCA followed two years after the plan's approval. In 1976, nineteen engineers traveled to the United States for training. They returned to Hsinchu to build a demonstration factory intended to teach manufacturing as well as produce chips. Government leaders treated mastery of the technology as a national goal.

A factory that taught Taiwan to make chips

Taiwan opened its first integrated circuit demonstration plant on October 29, 1977. The factory processed silicon wafers measuring three inches across, using a seven micrometer manufacturing process. Its intended products included chips for digital watches and other consumer devices.

A wafer is the thin slice of silicon on which many chips are made together. The factory combined a production line with a practical classroom: engineers had to operate the equipment, identify faults and find ways to make manufacturing more dependable.

Early setbacks were frequent. Within six months, however, the team raised its yield to about 70 percent, surpassing the original RCA line. Yield is the share of chips produced that work properly. A higher yield means more usable chips from each wafer, reducing the cost of each good chip and making profitable production more achievable.

The improvement showed that Taiwan's engineers could do more than follow instructions from abroad. They learned how to adjust and improve the process under local conditions. That habit of solving problems on the factory floor became a central part of the island's manufacturing culture.

CMOS uses complementary pairs of transistors, the tiny electronic switches inside a chip, that consume very little power when idle. In the 1970s, this made it well suited to watches and calculators running on small batteries. The choice connected a technical advantage to products that consumers already wanted.

The seven micrometer process described the size of features printed on the chip. At the time, it was advanced and difficult to master. Today's leading chips use process designations measured in nanometers, at scales thousands of times smaller. The basic work needed to achieve good yields still depends on clean rooms, precise lithography, stable chemistry and strict control of manufacturing conditions. Lithography is the step that transfers the chip's patterns onto the wafer.

The demonstration plant produced chips for Taiwan's first electronic watches. Exports grew, and Taiwan quickly became the world's third largest producer of electronic watches, with watch exports providing an early market for its new capability. The engineers had also gained experience building, debugging and improving a modern wafer production line. That knowledge was ready to move beyond a single factory.

Hsinchu gave new companies room to grow

The government established Hsinchu Science Park in 1979 to bring research groups, startups and suppliers together near universities and the ITRI campus. It provided ready land, a simpler setup process and shared utilities, alongside a community where engineers could exchange ideas and solve problems.

Engineers trained at the demonstration plant could form teams, seek investment and start companies close to tools and mentors. The park also attracted people returning from overseas and multinational companies looking for dependable partners. Research, manufacturing and business development could take place within the same industrial neighborhood.

ITRI spun off UMC in 1980, creating Taiwan's first commercial chip company. The state and local banks supplied its initial funding, and private investors were brought in afterward. The move gave knowledge developed through a public project a route into commercial production.

Alumni of the demonstration plant went on to establish or join businesses in chip design, testing and packaging. Companies such as Winbond and ASE expanded the network. Taiwan developed the ability to design chips, manufacture wafers, check finished chips, assemble them into packages and ship completed products.

That concentration of related businesses mattered alongside the growth of individual manufacturers. A wafer factory needs suppliers, trained workers and companies that can carry out the stages before and after fabrication. Hsinchu helped those activities grow together, making the industry more than a collection of isolated plants.

Morris Chang changed what a chip factory could be

ITRI launched a very large scale integration project in the mid 1980s to prepare for more complex chips. Very large scale integration means placing many electronic components together on a chip. The project sought to move Taiwan beyond its early consumer products into more demanding manufacturing.

Morris Chang, ITRI's new leader, drew on decades of industry experience to propose a different business model. Many chip designers needed a manufacturer but could not justify owning an expensive factory. Chang proposed a pure play foundry: a company dedicated to making chips for customers, without designing its own competing chips.

TSMC opened in Hsinchu in 1987. Philips invested about 27.5 percent, and Taiwan's government provided a decisive stake. The company offered designers a manufacturing partner whose business was to build their products, rather than compete with them through its own chip designs.

The model lowered the financial barrier for design companies. A business could concentrate on creating a chip and place its manufacturing order with TSMC. Such designers are often called fabless companies because they operate without their own fabrication plant.

TSMC could combine orders from many customers to build production scale. More volume supported investment in better processes, tighter cost control and greater reliability. Those improvements attracted more customers, whose orders helped fund the next round of development. The model allowed TSMC to outperform many integrated device manufacturers, businesses that both design and make chips.

How the early milestones led to today's industry

Taiwan's rise unfolded through several distinct stages. Government support established the first technical capability; a demonstration line trained engineers; new companies brought that capability into the market; and TSMC's foundry model widened the customer base.

• February 1974: Seven officials and engineers agreed at a Taipei breakfast meeting on a NT$400 million integrated circuit program.
• August 1974: Pan Wen-yuan's plan won approval.
• 1976: An RCA technology transfer contract supported the program, and nineteen engineers traveled to the United States for training.
• October 29, 1977: Taiwan opened its first integrated circuit demonstration plant.
• 1979: The government established Hsinchu Science Park.
• 1980: ITRI spun off UMC as Taiwan's first commercial chip company.
• Mid 1980s: ITRI pursued very large scale integration under Morris Chang's leadership.
• 1987: TSMC opened in Hsinchu with investment from Philips and the government.
• Early 2020s: TSMC produced the majority of the world's contract chips and most of its advanced processors.
• 2025: Tariff changes, rules on chip origin and production in Arizona brought new pressures and opportunities.

By the early 2020s, TSMC was manufacturing advanced processors used in smartphones, graphics cards and artificial intelligence servers. It was producing chips in volume using three nanometer technology and preparing two nanometer technology.

Close relationships with critical equipment suppliers, including the maker of extreme ultraviolet lithography tools, became a strategic advantage. These tools help print the extremely fine patterns needed for advanced chips. TSMC's focus on manufacturing for other companies connected that technical progress to customers across the global electronics industry.

A network built through contracts and training

Taiwan acquired its initial capability through a formal technology transfer agreement, overseas training and the work of engineers who brought their knowledge home. The RCA contract was a sale. The Technology Advisory Council connected the project to expertise among people living abroad.

The demonstration line gave Taiwan a place to put that knowledge into practice. Spinoffs and the science park then spread it through businesses that welcomed foreign partners. The resulting industry connected designs from California, equipment from Europe and manufacturing experience developed in Hsinchu.

That history remains relevant to two debates: how quickly advanced manufacturing can be recreated elsewhere, and whether Taiwan will surrender the strategic value often called its silicon shield. The term describes the importance attached to Taiwan's chip industry in discussions of its security and international position.

Replicating the island's industrial cluster requires money, time and people trained across many specialties. Honghong Tinn, an assistant professor who studies the history of Taiwan's technology industries, described the cost challenge of expanding chip manufacturing in the United States:

Making chips at scale in the U.S. is obviously going to be more expensive, so they would probably need huge subsidies.

Jonas Parello-Plesner, a policy analyst who has studied technology strategy, described Taipei's determination to retain advanced capacity at home:

There is no way Taipei is relinquishing its silicon shield.

The two observations address different parts of the same manufacturing system. New factories require financial support and skilled teams, while the location of advanced capacity also carries strategic value for Taiwan.

Tariffs put the location of manufacturing in focus

A new round of United States tariffs in 2025 created uncertainty for Taiwan's exporters. Semiconductors received temporary relief, but companies continued to face pressure to move some production and tighten sales controls.

Trade friction between the United States and China had already redirected investment and supply chains. Some companies moved final assembly away from mainland China. Taiwan's businesses, selling into both markets, had to weigh costs, reliable production and access to customers.

China then clarified its approach to the origin of chips for customs purposes. Under that approach, origin depends on where the wafer is fabricated, regardless of where the chip is designed or packaged.

A chip can pass through several countries before reaching a customer. The Chinese approach puts the deciding weight on the factory that makes the wafer, rather than the location of the design team or the company completing the package.

Chinese authorities treat Taiwan as domestic for this purpose. Chips fabricated in Taiwan therefore avoid the punitive tariffs China applies to many United States goods. The approach disadvantages companies fabricating in the United States and favors foundries in Taiwan and mainland China.

For manufacturers and customers, the customs treatment adds another consideration when choosing a factory. Alongside price, technical capability and dependable delivery, the location of wafer fabrication can affect the tariff attached to the finished chip.

Arizona shows why one factory is only part of the chain

TSMC continued building its manufacturing campus in Arizona as trade pressures grew. In 2025, Nvidia unveiled its first wafer made in the United States for its Blackwell generation at the site.

The wafer still needed advanced packaging, a complex production stage that TSMC carries out at scale in Taiwan. One of its techniques is chip on wafer on substrate, which brings chips together into a package that connects them for use in a finished system.

A packaging partnership with a United States company specializing in outsourced assembly and testing was also in progress, with operations expected to start in the middle of the decade. The Arizona wafer and the packaging work illustrate how production extends beyond the fabrication plant.

The manufacturing flow described in 2025 combined wafer fabrication in the United States with advanced packaging in Taiwan. Expanding fabrication abroad therefore involved connecting a new factory to other specialized stages of production.

Taiwan's experience helps explain the scale of that task. Its chip industry grew through linked investments in engineers, factories, suppliers, design businesses, testing and packaging. Building a plant addresses one part of that network; developing the surrounding capabilities requires further work.

What the breakfast decision changed

The 1974 plan succeeded by connecting a national economic goal to specific steps. Leaders chose CMOS for a market Taiwan could serve, used a demonstration line to develop practical skills and supported a science park where talent, suppliers and investment could gather.

Engineers were trusted to solve difficult manufacturing problems, and new firms carried their experience into commercial markets. Government ownership did not remain unchanged: public stakes were sold down as private investors gained confidence, while the industry's own research spending increased as it matured.

Advanced manufacturing now demands substantial amounts of power and water, along with specialized skills. Those requirements add to the cost and complexity of expanding production, whether in Taiwan or abroad.

The lasting contribution of the Taipei breakfast was the system built after it. A decision to import technology became a program for learning it, improving it and sharing it through new businesses. Patient investment, a closely connected industrial neighborhood and engineers learning through production helped turn an economic crisis into a new source of prosperity.

What to Know

• A breakfast meeting in February 1974 led to a NT$400 million program to develop integrated circuit manufacturing in Taiwan.
• ITRI and ERSO managed the RCA technology transfer, and nineteen engineers trained in the United States in 1976.
• The demonstration plant opened on October 29, 1977, using wafers measuring three inches across and a seven micrometer process. Yields reached about 70 percent within six months.
• Watch chips supplied an early market as Taiwan became the world's third largest producer and exporter of electronic watches.
• Hsinchu Science Park was established in 1979, and UMC spun off from ITRI in 1980.
• TSMC opened in 1987 under Morris Chang's foundry model, with Philips investing about 27.5 percent and the government supplying a decisive stake.
• By the early 2020s, TSMC led contract manufacturing and advanced processors, producing three nanometer chips and preparing two nanometer technology.
• Tariffs and China's rules on wafer fabrication placed new weight on where chips are manufactured.
• TSMC's Arizona expansion brought wafer production to the United States, while advanced packaging remained concentrated in Taiwan.
• Taiwan's industry grew through formal technology transfer, overseas expertise, practical training, connected businesses and continued investment.