In the past few years, Europe has once again prioritized the development of semiconductors.

In February 2022, the European Commission proposed a comprehensive set of measures aimed at strengthening the EU's semiconductor ecosystem, known as the so-called "European Chips Act." At that time, the global shortage of semiconductors and the crazy rise in chip prices also made European semiconductors one of the focuses of the industry.

The European Chips Act clearly states that the EU's goal is to increase semiconductor production capacity to 20%, and plans to invest 43 billion euros to support the development of chips in the future. More than two-thirds of the funds are designated for the construction of new, leading chip manufacturing plants, or "large-scale wafer factories."

Advertisement

With the official implementation of this act, semiconductor companies have begun to invest in the construction of new wafer factories. Infineon, STMicroelectronics, Bosch, Intel, Wolfspeed, TSMC, and other large factories have joined in, and the revival of the European semiconductor industry seems to be no longer a luxury.

Investment requires the right timing, location, and people. The right timing is naturally the increasingly prominent international status of semiconductors today, with countries constantly competing around it, and the EU has also given a series of subsidies and preferential policies; the right location is Europe's mature supporting infrastructure and suppliers, coupled with a huge automotive market, to protect the new wafer factories; the right people are the talents, Europe has a high level of education, and there should be no shortage of talents.

On August 8, 2023, TSMC officially announced a joint investment with Bosch, Infineon, and NXP in the European Semiconductor Manufacturing Company (ESMC). This wafer factory is located in Dresden, Germany, and plans to start construction in the second half of 2024, with the goal of starting production by the end of 2027. It is expected to have a monthly production capacity of 40,000 12-inch wafers, providing TSMC's 28/22-nanometer planar CMOS process and 16/12-nanometer FinFET process. Among them, TSMC will hold 70% of the joint venture and be responsible for operations, while Bosch, Infineon, and NXP will each hold 10%, becoming a representative of the newly built wafer factories in Europe.

However, many people are not optimistic about the prospects of this wafer factory, and some even strongly oppose it. It is important to know that the EU has been looking forward to TSMC for two or three years, but why has it encountered internal doubts when it comes to building a factory?

Behind the prosperity

The first thing we need to make clear is that the wafer factory is not a steel mill. It cannot produce chips like steel billets with full effort. The use of steel billets is very broad, but the vast majority of chips can only meet the needs of specific products. The most important thing is how much of the theoretical production capacity can be filled with orders. Even TSMC, which is in short supply during the peak season, also needs to seek new customers in the off-season to utilize the surplus production capacity.So, how many customers can ESMC find in Europe?

The answer may be different from what many people think. The main customers of the wafer factory are, of course, the three companies that participated in the financing, namely Bosch, Infineon, and NXP. Later, it may also include one of the three giants, STMicroelectronics. These four companies are the main customers of ESMC in Europe.

It sounds quite a lot, but the problem is that these four companies are not like fabless companies such as NVIDIA, Qualcomm, and AMD, but old-line IDM manufacturers. They each have their own wafer factories. Although TSMC's process technology is more advanced and mature, the relative cost is also higher. Once the wafer factory is built, it may not become the first choice for these giants.

Some people may ask, what about Europe's fabless companies? Can't they provide orders for ESMC?

It's really a bit uncertain. According to a previous report, Europe's share of the global fabless chip sales was only 2% in 2015, down from 4% in 2010. The reason is that many companies and business departments were acquired by American companies, such as CSR, headquartered in the UK, which is the second-largest fabless integrated circuit supplier in Europe and was acquired by Qualcomm in the first quarter of 2015; Lantiq, headquartered in Germany, is the third-largest fabless integrated circuit supplier in Europe and was acquired by Intel in the second quarter of 2015.

At that time, the largest European fabless company, Dialog, achieved sales of $1.2 billion in 2016 and was the only company in the fabless TOP 50 at that time with its headquarters in Europe. As a European old-line chip company listed in Frankfurt, Germany, Dialog mainly provides mixed-signal chips for the consumer electronics field, including power management chips, lighting chips, Bluetooth chips, audio subsystems, etc.

It is worth noting that Dialog has been supplying power management chips for Apple's iPhone since 2007, and Apple's orders once accounted for more than 70% of Dialog's annual revenue. In 2018, Apple announced the acquisition of part of Dialog's assets for $600 million, including about 300 engineers' R&D team and some equipment technology, as well as Dialog's offices in the UK, Italy, and Germany. Dialog began to withdraw from the field of mobile power management chips.

Subsequently, Dialog did not collapse due to Apple's acquisition and began to layout the automotive and smart home fields. In 2019, Dialog acquired Creative Chips, an industrial Internet of Things IC supplier, for $80 million, and acquired Silicon Motion's mobile RF IC business for $45 million; in 2020, Dialog acquired Adesto Technologies, a supplier of integrated circuits (ICs) and embedded systems, for $500 million.

Time came to 2020, although Dialog experienced a certain decline due to the withdrawal of Apple's orders, it is still one of the representatives of Europe's fabless and can even maintain the tenth place in the global chip design company.

However, such brilliance did not last long. At the beginning of 2021, Japan's Renesas Electronics acquired Dialog for $5.9 billion, and Dialog became a Japanese company from a British company. So far, the former Fabless three strong of Europe have all been taken over by foreign companies.Today, the only identifiable European fabless semiconductor company is Nordic, headquartered in Norway. This company is a wireless IoT system SoC manufacturer, whose low-power Bluetooth solutions are widely used in gaming, sports, fitness, mobile accessories, consumer electronics, and other IoT fields. Its total revenue in 2022 was $776 million, ranking 39th among global fabless semiconductor companies, and it is also the only European fabless semiconductor company in the Top 100.

Looking at the present, the European fabless semiconductor industry has long been forgotten by many people, accounting for less than 1% of the global fabless semiconductor revenue. Compared to Japan and South Korea, the situation faced by European chip design is obviously much more severe, which puts ESMC in an awkward position. Setting up factories in Europe may not necessarily serve European customers, and apart from thousands of manufacturing jobs, it may not bring many immediate benefits.

Light design, heavy manufacturing

Looking back, the European Chips Act seems to support the development of the chip industry, but it is not a comprehensive view, but slightly "biased" to tilt subsidies towards wafer fabs such as ESMC and Intel Madgeburg, while fabless companies have not gained much benefit in this grand development activity.

Nordic's Chief Technology Officer, Svein-Egil Nielsen, said, "The EU Chips Act focuses heavily on capital expenditure, but fabless companies like us have not been given many opportunities."

Nordic's Chief Operating Officer, Svenn-Tore Larsen, said, "The act proposes many measures, but the new wafer fab is not for mixed-signal processes, so we must stick to using existing suppliers for a considerable period of time."

When European media are vigorously promoting TSMC and Intel's establishment in Germany, fabless companies like Nordic are ignored. A super-large project with an investment of tens of billions of dollars easily crushes a company with an annual revenue of hundreds of millions of dollars. This is the harsh reality they face.

It is quite tragic that the neglect of chip design has occurred more than once in history.

In 2013, the EU determined the need to strengthen chip design capabilities and the fabless industry. The EU and the semiconductor industry began to plan "to strengthen its electronic design industry and fabless semiconductor companies"; in 2018, the EU once again planned to strengthen the EU chip design ecosystem through the "European Design Alliance" and the "Strategic Design Initiative"...

Ten years have passed, and the European fabless industry has not only not developed but has shrunk by 50%. The EU's latest ten-year plan, the "2030 Digital Compass," focuses entirely on manufacturing, with little clear vision for investment in the chip design industry. No wonder the two senior executives of Nordic are so critical.On the contrary, Apple on the other side of the ocean is more attentive to European chip design. It announced an investment of 1 billion euros in 2021 to establish a new chip design center in Munich, Germany, and then added another 1 billion euros in investment in 2023. Just this 2 billion euros has exceeded the total investment of the European Union in the field of chip design in the past 10 years.

Of course, no matter what, the EU is still a bit better than the neighboring UK. Arm and Dialog have all gone abroad, and the UK government has only just announced a semiconductor support plan in a panic, with a total amount of 1 billion pounds (about 1.26 billion US dollars), spread over ten years, with an initial investment of 200 million pounds (about 250 million US dollars) from 2023 to 2025, to support domestic semiconductor development.

Everyone is drawing a big picture, but the EU's picture is a bit bigger, and the local chip manufacturing industry has more or less benefited. The UK's picture, whether it is manufacturing or design, does not benefit much.

For the matter of focusing on light design and heavy manufacturing, the European semiconductor industry is not without clear-headed people.

German think tank Stiftung Neue Verantwortung (SNV) analyst Kleinhans said in an interview with Reuters in 2021: "The problem with the EU strategy is that, unlike the US and Asia, Europe lacks a meaningful chip design industry, which cannot justify the cost of large wafer factories."

"In terms of local chip capacity, it is not enough to fill a wafer factory at all," he said, "This means that EU wafer factories need to attract foreign customers - this is extremely difficult."

SNV also released a 30-page report in 2023 titled "European Semiconductor Manufacturing Shortage - Why 2nm Wafer Factories Are Not Good Investments?", concluding that the EU's continued focus on cutting-edge chip manufacturing to strengthen its technological sovereignty in the semiconductor field is unwise and may waste tens of billions of dollars.

The report pointed out that although the EU lacks cutting-edge wafer factories, more importantly, it lacks the design capabilities of advanced logic chips and fabless industries. Increasing supply (wafer factories) before there is demand (fabless companies) within the EU is, at best, overly optimistic and, at worst, naive.

SNV said that the EU has world-class RTOs, equipment suppliers, and silicon wafer suppliers, all of which are deeply involved in cutting-edge chip manufacturing in South Korea, Taiwan, and the United States. However, Europe's own semiconductor industry is mainly in the design of logic chips at cutting-edge and mature process nodes for industrial and automotive applications. Therefore, if the EU is to invest in wafer factories today, these investments should focus on wafer manufacturing at 14 nm and above.

The initiatives of the wise are obviously not taken to heart by the EU. Two years later, TSMC's leading large wafer foundry still settled in Germany, and the question of how much return can be obtained by investing billions of dollars in building a factory is destined to remain unanswered.The Overlooked Other Industries

As the saying goes, the squeaky wheel gets the grease. This phrase couldn't be more fitting for the European semiconductor industry. Whoever has the loudest voice or the largest scale can reap more benefits from the chip act, and whoever can demonstrate the most direct economic benefits can receive the most subsidies.

Chip design has been neglected due to its small proportion, and the industries surrounding chip manufacturing are not faring much better.

Europe has a strong advantage in producing highly customized complex materials and chemicals used for advanced chip manufacturing. These products, in addition to being frequently mentioned by Japanese companies, also come from a few European companies such as Merck, BASF, and Solvay.

Their current situation in Europe is not optimistic. In fact, compared to the United States and Taiwan, these chemical and material suppliers have a much smaller scale of investment. Some industry insiders believe that one reason is that the European chip act does not fully cover investments beyond chip manufacturing, and European environmental regulations make it more difficult to expand chemical facilities. Coupled with the recent natural gas crisis, which has driven up already expensive energy prices, forcing manufacturers to shut down or suspend production of some products. Industry executives also said that without strong incentives, continuing to expand in Europe is not a good choice.

"Supplying chemicals to new semiconductor factories requires investment in dedicated assets. Therefore, the lack of national support will definitely become an obstacle for chemical suppliers," Solvay's President Rodrigo Elizondo told the Financial Times. "In our view, the lack of a strong regional chemical supply will definitely endanger the operation of European semiconductor factories."

BASF Senior Vice President Lothar Laupichler said, "Everyone is talking about semiconductor manufacturing, but there is not enough attention to the chemicals needed to produce these chips. People almost feel that chemicals are like water or electricity, and they can come out directly by turning on the tap, but this is a misunderstanding."

Merck Executive Board member and CEO of the Electronics Division Kai Beckmann added, "We need to work with the EU on this issue because we are talking about very highly specialized materials, and these materials may not be well reflected in Europe's ambitions."

These three companies predict that unless the EU invests in the field of materials and chemicals, there may be a shortage of local supply in Europe when more wafer factories are completed, and it will have to be imported from other regions.

In addition to materials, Europe is also facing a shortage of skilled workers. Although Europe has top research institutions such as IMEC and CEA-Leti, and a number of higher education institutions, they seem to be powerless to solve the shortage of skilled workers. According to a survey by the European Labor Agency on the largest labor shortage in the EU, engineers and technical personnel have become one of the main talent gaps in 10 countries.Companies such as Infineon Technologies from Germany, Edwards Vacuum from the UK (a key component and subsystem supplier for ASML), and Austria's AT&S (one of the main suppliers for high-end substrates that install semiconductors) have warned that foreign talent is crucial for the further development and sustainability of the European semiconductor industry.

AT&S CEO Andreas Gerstenmaye stated that the company is striving to find the 800 skilled workers needed for its new R&D center in Austria, "We must recruit talent globally because there is no experience and technology here (in chip substrates)."

Infineon's Head of Human Resources, Martin Stöckl, said that the entire supply chain will be chasing the same talent, which exacerbates the situation. "The talent shortage in Europe is very serious," he said, "If you quickly calculate, we (Infineon) will build a new wafer factory, STMicroelectronics and Intel are also expanding. In the next few years, we will need at least thousands of engineers and technical personnel."

Chip design, chip materials, chip talent, once the strengths of Europe, have now become weaknesses. These issues have existed for more than a decade or even more than twenty years. If there were no geopolitical influences, European semiconductors might fall into a slow decline. However, the introduction of the European Chips Act is like a strong dose of medicine for a critically ill patient. Whether it is a flash in the pan or a successful rescue can only be left to fate.

In conclusion, why does Europe lack advanced process wafer factories? The main reason is that Europe lacks the demand for advanced processes. The three giants are all experts in power chips and rarely venture into logic chips, and the decline of fabless companies has also made it difficult for European foundries to find a foothold.

When we delve deeper into the reasons, we will find that the decline of Europe's consumer electronics industry is the more fundamental cause of this phenomenon. The exit of Nokia, Ericsson, Siemens, and Alcatel from the mobile phone industry has dealt a heavy blow to Europe's related chip industry. Without the demand for local chip companies, they are like rootless duckweed. Coupled with frequent mergers and acquisitions of large companies, European chips have finally retreated to the automotive field.

Understanding this, and then looking at the factory construction boom brought about by the Chip Act, how much difference is there from seeking fish by climbing a tree?