Published: March 10, 2026
In the previous volume, I discussed foundries that specialize in front-end contract manufacturing. In Japan, “contract manufacturing” is often associated with the term “subcontracting.” With that said, the subject of this article is, “Is a foundry just a subcontractor?” In the final section, I will discuss OSAT (outsourced semiconductor assembly and test), which specializes in back-end process and inspection, in relation to subcontracting.
Table of Contents
Two Types of Foundries
Is a Foundry Just a Subcontractor?
OSAT (Outsourced Semiconductor Assembly and Test)
Volume 16: Horizontal Specialization in the Semiconductor Industry—Is a Foundry Just a Subcontractor?
Two Types of Foundries
Before discussing the following, I would like to explain that there are roughly two types of foundries or contract manufacturing of front-end process. It is fair to say that these two types of contract manufacturing are completely different businesses. Our company has experience with both types.
One type is the contract manufacturing of products made at the factories of integrated device manufacturers (IDMs). This occurs when demand exceeds production capacity and IDM chooses to outsource manufacturing to companies such as foundries instead of investing in increasing its own production capacity. It is a way to respond to fluctuating demand by increasing variable costs rather than fixed costs. In some cases, foundries’ and other contractors’ wafer processes can be used as is or with only minor modifications. In other cases, however, a new wafer process identical to the IDM process must be developed.
Note: In Volume 2, I explained, “The manufacturing process for fabricating elements such as transistors and integrated circuits in the surface of a silicon wafer is generally referred to as a wafer process or simply a process in the industry.” In this section, I use the term “wafer process” instead of “manufacturing process.”
The other type is designing a product based on a wafer process developed by a foundry. In the assembly industry, I imagine that the products come first, and the manufacturing process is designed accordingly. However, in the case of semiconductors, especially integrated circuits, the front-end or wafer process comes first. Initially, foundries develop wafer processes for manufacturing transistors and other elements. Customers who use foundries, such as the design departments at IDMs and fabless semiconductor companies, play a role in designing circuits and systems using those transistors and elements. Product design can only begin after the wafer process has been established.
Foundries provide not only wafer processes but also the necessary information, data, tools, and design assets (known as IP in the semiconductor industry) to design products in cooperation with related companies. In order for foundries to undertake manufacturing, it is essential that users design their own products. Therefore, foundries must provide the necessary tools and environment to support their users. I think they are more active in that they provide a platform for designing and manufacturing products rather than being a passive contract manufacturing business.
Note: IP originally stands for intellectual property. In the semiconductor industry, however, IP refers to design information for various functional blocks. Because it is difficult to design all circuits from scratch in-house for a large-scale product, IP is often combined with in-house designed circuits to achieve the desired functions.
Is a Foundry Just a Subcontractor?
Although I didn’t watch it when it aired, there was a Japanese TV program called Shima Kosaku’s Asian Entrepreneurs. The second episode, which aired in June 2013, was titled “Morris Chang – A Subcontractor Changed the World.”
In Japan, people tend to associate contract manufacturing—the business of manufacturing products for other companies—with the Japanese word “shitauke,” which means subcontracting in English. Aside from the formal definition of subcontracting, the term subcontracting tends to imply that an outsourcing party is in a higher position and a contractor, a foundry, is in a lower position. This is just a feel from my experience, I would say that there used to be such an atmosphere in the Japanese semiconductor industry around 2000.
Around 2000, when I visited a foundry in Southeast Asia, I heard the following story. A Japanese client company had visited the foundry, and they acted similarly to the character in the illustration below. The foundry employees were businesslike during the meeting. However, after the clients left, the employees complained, “How can we work so hard for clients like that?” Would anyone want to work hard for a customer who treats them that way? Of course, we don’t do that.
Image of subcontracting
In the case of the first type of contract discussed in the previous section, i.e., if the company was only contracted to manufacture IDM’s products, the term “subcontractor” might apply. As mentioned in the previous volume, TSMC also appeared to be primarily involved in this business in the early days. However, in the case of the second type described in the previous section, although these companies specialize in front-end contract manufacturing, they are not just contract manufacturing companies. As mentioned in the previous section, they provide fabless companies with everything they need to design their products in collaboration with other companies. In other words, they offer a product design platform. Conversely, the foundry does everything except product design. They are beyond contract manufacturing. They are not just subcontractors.
Specialization in the automotive industry is structured like a pyramid, with finished car manufacturers at the top. I can’t help but view this structure as hierarchical. (I apologize if I’m wrong.) I don’t think the relationship between fabless companies and foundries is hierarchical. Rather, it’s a win-win relationship, or an interdependence of equals. The two are inseparable. Without fabless companies, foundries would not exist. Likewise, without foundries, fabless companies would not exist.
OSAT (Outsourced Semiconductor Assembly and Test)
In Volume 13, I introduced the three types of players involved in horizontal specialization: fabless companies, which specialize in design; foundries, which specialize in front-end processes; and OSATs, which specialize in back-end processes. I discussed fabless companies in Volume 14 and foundries in Volume 15. However, I had not yet mentioned OSATs. Here, I will briefly discuss OSATs while also keeping in mind the term “subcontractor.”
I tried to find out when the term “OSAT” started being used for companies that specialize in back-end and inspection processes, but I couldn’t find the answer. Perhaps this is because I originally had a background in front-end, but I think I first heard the term around 2010. Maybe it was a little later. Prior to that time, I recall many people in Japan using the term “subcon.” “Subcon” is short for “subcontractor.” Using the term “OSAT” instead of “subcon” suggests that back-end manufacturing contractors are more than just subcontractors. I think it can be said that they have become more than subcontractors by becoming active providers of unique services and solutions, rather than simply passive contractors who are entrusted with a portion of a contractor’s work.
Note: subcontractor: a person or company that does part of a job that another person or company is responsible for (Cambridge dictionary)
As I discussed in the previous volume, separating design from manufacturing became a major theme for front-end processes. However, back-end processes are less tied to design and easier to separate. They are also less tied to the front-end process and easier to separate. Additionally, the technology level is not so high and factory construction costs are relatively low compared to the front-end process, so the barriers to entry are low. Furthermore, unlike front-end processes, back-end processes were highly labor-intensive in the early stages, and their labor costs seemed to account for a high percentage of the cost.
These factors, along with the expansion of consumer applications and increased cost competition accelerated the shift of back-end processes to Asia, where labor costs were low, as early as the 1960s. In this trend, some local companies entered the back-end contract manufacturing business or founded companies that specialize in back-end contract manufacturing, in addition to business forming local subsidiaries or joint ventures with local companies. These companies are now called OSATs. At that time, however, I think they were probably undertaking the back-end processes of IDM products based on the technology they had absorbed from the IDMs. “Subcon” would probably still be the more appropriate term.
Here are a few examples of OSATs that were established from the late 1960s to the early 1970s. Among the leading OSATs that are still in existence as of 2025, Amkor (U.S.), the second-ranked company by revenue, was founded in 1968. Initially, Amkor was a South Korean company called ANAM, and its U.S. base was Amkor. Founded in 1971, Orient Semiconductor Electronics (OSE) of Taiwan continues to operate today as a leading OSAT. According to its website, OSE is “the very first Taiwan-funded OSAT” (as of May 2025) and appears to be the first OSAT in Taiwan. Another leading OSAT, Carsem of Malaysia, was founded in 1972. It appears that the company was founded by the former manager of Fairchild Semiconductor’s back-end plant in Hong Kong. All of these companies predate TSMC, which was founded in 1987.
Both ASE (Taiwan) and SPIL (Taiwan) were established in 1984. They are ranked first and fourth by revenue, respectively. Although they were founded much later than the aforementioned three companies, they predate TSMC, which was founded in 1987.
Note: ASE and SPIL merged in 2018. They operate as separate companies under a holding company called ASE Technology Holding. TrendForce’s 2021 ranking shows the rankings of each ASE and SPIL company, not the holding company of ASE.
Note: Each OSAT ranking is based on 2021 data from TrendForce. As far as I researched, TrendForce has not published any data beyond 2022 as of May 2, 2025. The most recent data available is IDC’s 2022 ranking. Based on this information, ASE Technology Holding and Amkor are ranked first and second, respectively. The respective rankings of ASE and SPIL are not shown.
Note: As mentioned above, although Amkor is now a U.S. company, all of its manufacturing bases are in Asia. There are also manufacturing sites in Japan that were originally back-end process plants belonging to Japanese IDMs.
Lastly, I would like to add the OSAT-related events described here to the chronology of fabless companies and foundries from the previous volume.
| Year | Fabless Companies | Foundries | OSATs |
| 1960s | IDMs expanded their back-end processes to Asia, where labor costs are lower. | ||
| 1965 | Advocacy of Moore’s Law (U.S.) | ||
| 1968 | Foundation of ANAM (South Korea), predecessor of Amkor (U.S.) | ||
| 1969 | Foundation of LSI Computer Systems Inc. (US) | ||
| 1971 | Foundation of OSE (Taiwan) | ||
| 1972 | Foundation of Carsem (Malaysia) | ||
| 1973 | Foundation of ITRI (Taiwan) | ||
| 1974 | Foundation of ERSO in ITRI (Taiwan) | ||
| 1979 | Professor Mead’s advocacy of “separation of design from manufacturing” (U.S.). | ||
| 1980 | Foundation of UMC (Taiwan) Note: At this time as IDM |
||
| Publication of Introduction to VLSI Systems by Mead and Conway (U.S.) | |||
| 1981 | Mead and Conway received Electronics Magazine’s Achievement Award. (U.S.) | ||
| 1983 | Foundation of Altera (U.S.) | ||
| 1984 | Foundation of C&T, Xilinx, and Cirrus Logic (U.S.) | Foundation of ASE (Taiwan) and SPIL (Taiwan) | |
| 1985 | Foundation of Qualcomm (U.S.) | ITRI Taiwan appointed Morris Chang as its president. (Taiwan) | |
| 1987 | Foundation of TSMC, a pure-play foundry company (Taiwan), the world’s first foundry | ||
| 1991 | Foundation of Broadcom (U.S.) | ||
| 1993 | Foundation of NVIDIA (U.S.) | ||
| 1994 | Foundation of the Fabless Semiconductor Association (FSA) | ||
| 1995 | UMC changed its business model from IDM to a pure-play foundry. (Taiwan) | ||
| 1997 | Foundation of MediaTek (Taiwan) | ||
| 2007 | FSA was reorganized into the Global Semiconductor Alliance (GSA). | ||
| 2009 | AMD changed its business model to a fabless company. (U.S.) | Foundation of GlobalFoundries (U.S.) | |
| 2018 | Business integration between ASE (Taiwan) and SPIL (Taiwan) | ||
| 2021 | Intel announced the strengthening of its foundry business. (U.S.) | ||
| 2022 | Intel announced the acquisition of Tower. (U.S.) | ||
| Professor Emeritus Mead received the Kyoto Prize. | |||
| 2023 | Intel announced the termination of the Tower acquisition. | ||
| 2024 | Intel announced that it will spin off its foundry business into a subsidiary. | ||
As mentioned above, OSATs emerged more than a decade before fabless companies and foundries. I think that OSATs were essentially subcontractors of IDMs at the beginning stage, but they gradually gained competence and transformed from just subcontractors into OSATs.
This was supposed to be the final post in my series on horizontal specialization. However, as I write this, the volume has grown again. The next post will be the last one in the series and will discuss the fact that horizontal specialization does not advance in all products or countries.
Click below to read this series.
Semiconductor Miniaturization:
Volume 1: Semiconductor Miniaturization: What is Moore’s Law?
Volume 2: Semiconductor Miniaturization and Manufacturing Process
Volume 3: Semiconductor Miniaturization and International Technology Roadmap
Volume 4: Semiconductor Miniaturization and Semiconductor Business
Volume 5: Semiconductor Miniaturization and Semiconductor Business (Part 2)
Volume 6: Semiconductor Miniaturization and Semiconductor Devices
Volume 7: Semiconductor Miniaturization: What is MOSFET Scaling?
Volume 8: Semiconductor Miniaturization: Limitations of MOSFET Scaling
Volume 9: Semiconductor Miniaturization and Analog Circuits
Shift to Larger Diameter Silicon Wafers:
Volume 10: Shift to Larger Diameter Silicon Wafers: How a Common Material, Silicon, Became a Main Player
Volume 11: Shift to Larger Diameter Silicon Wafers (Part 2): How Silicon Wafers Are Made
Volume 12: Shift to Larger Diameter Silicon Wafers (Part 3): Reasons and History
Horizontal Specialization in the Semiconductor Industry
Volume 13: Horizontal Specialization in the Semiconductor Industry and the Rise of Fabless Companies
Volume 14: History of Horizontal Specialization in the Semiconductor Industry–Emergence of Fabless Semiconductor Companies
Volume 15: History of Horizontal Specialization in the Semiconductor Industry–Emergence of Foundries
Volume 16: Horizontal Specialization in the Semiconductor Industry—Is a Foundry Just a Subcontractor?

Comment