China is the world's largest chip consumer market, but because the United States prohibits all companies that use American technology and equipment from shipping to the Chinese market, high-end lithography machines used to produce chips are restricted from entering China. Almost every once in a while, policies related to whether the world's high-end lithography machines can be introduced into China will become hot news.
As Cai Hongping, a well-known hard technology investor, said: "Since the first day of semiconductor research, it has been globalized. No country can cover (the entire industrial chain), nor can the United States." Therefore, it is as strong as the United States, and it cannot produce The production of top lithography machines can only be restricted by other countries through hooliganism.
picture
Image source | ASML
Compared with big countries like the United States, China, Germany and Japan, the Netherlands is an out-and-out small country, but its local company ASML, which produces high-end lithography machines, has a high status in the semiconductor industry. This technology company located in Wildhofen, a small town in the south of the Netherlands, occupies more than 80% of the global high-end lithography machine market. It took only 31 years from 1984, when it was founded in the prefabricated house, to 2009, when it took the lead in the industry.
Understanding the rise of ASML, or understanding how the Netherlands cultivated ASML, should be of some value in solving the current bottleneck problem in China.
Where is the lithography machine more difficult than the atomic bomb?
Because the neck-stuck problem has been very prominent in the past two years, more and more ordinary Chinese people know about the lithography machine. But hearing the term lithography machine does not mean that you really understand it. Many people say that when we were so poor, we all produced the atomic bomb. Now that the economic situation is much better, are we still afraid that we will not be able to produce a lithography machine? People with a similar mentality, represented by Dong Mingzhu, once made a bold statement that she would use 50 billion to enter the semiconductor field.
Currently, the world's most advanced EUV lithography machine sells for as much as 120 million euros, which is comparable to a Boeing 737 airliner. The next-generation High-NA EUV lithography machine, which will be mass-produced in 2024, is expected to sell for between 300 million and 350 million euros. The cumulative R&D investment of EUV lithography machines exceeds 10 billion euros, and the threshold for players in the semiconductor industry is 100 billion yuan, and continuous investment is required. Blunting out 50 billion yuan to solve the problem exposed the unfamiliarity of the parties to this industry.
Both the atomic bomb and the lithography machine are difficult to manufacture. There may be only one key difference between the two, that is, after the atomic bomb is manufactured, the research project will end; but after the lithography machine is manufactured, the project has just begun, and enterprises must keep up with it. Demand, continuous optimization of performance while reducing costs is a typical "both must and must". Some investors have calculated that if atomic bombs are made according to the efficiency of photolithography machines, the price of an atomic bomb is only 100,000 yuan.
In short, the atomic bomb is a question of whether it is available, and the lithography machine is a question of whether it is good or not, which belong to two dimensions.
For TSMC, Samsung, or SMIC, when companies invest billions of dollars to build production lines, the most fearful thing is that equipment failure will cause production to stop. Once production is stopped, the loss will be huge. Therefore, this requires that the lithography machine must maintain absolute stability for 7x24 hours. This must rely on business methods to continuously maintain high profits-high R&D investment-continuous iterations, thus forming a cycle with a positive feedback effect.
An EUV lithography machine weighs 180 tons, has more than 100,000 parts, and requires 40 containers to transport. It integrates top technologies in multiple industries such as optics, organic chemistry, instrumentation, mechanical equipment, automation, and image recognition. It only needs to be installed and debugged It will take more than one year, and the annual downtime should not exceed 3%.
The manufacturing process of the reflective lens used in the EUV lithography machine is extremely complicated. In order to ensure that the light is not lost or deformed when it passes through the lens in a vacuum, the required technical precision is equivalent to laying an iron chain between Beijing and Shanghai, and the rails are undulating. No more than 1mm. And the imaging precision of these mirrors is so high that it is equivalent to taking a flashlight on the earth to shine on the moon, and the light spot is no more than the size of a coin.
It is precisely because of the super stability of the lithography machine that today, in the Shanghai Advanced Semiconductor Factory, there is still an ASML lithography machine imported 30 years ago that is still working around the clock. This is the earliest 5-inch chip production line in China. It was originally established in 1988 as a joint venture between Philips of the Netherlands and several state-owned semiconductor companies in Shanghai.
Philips not only helped mainland China keep up with the pace of international semiconductor development, but also had a great influence on the start of the semiconductor industry in Taiwan, which will be mentioned later. However, in terms of the greatest contribution of this industrial giant to the lithography machine industry, it is still necessary to return to the nurturing grace of ASML.
Backed by Philips, it is good to enjoy the shade under the big tree
According to the latest data (December 9, 2022), the market value of ASML is 242.4 billion U.S. dollars, and Philips is only 12.38 billion U.S. dollars, which is not even a fraction of ASML. What's more interesting is that Philips holds about 5.8% of ASML's shares, which translates to $14.06 billion, exceeding its own market value.
But if you turn the clock back to 1984, the situation is very different. In that year, Philips succumbed to cooperation with the Dutch Advanced Semiconductor Materials Company (ASM) to establish ASML. In addition to providing US$2.1 million in cash or equipment, 45 R&D engineers from Philips Natlab also joined ASML.
Although blessed by the aura of Philips Natlab, the company has not made a profit in the 10 years since its establishment. At the beginning of its establishment, no one cared about ASML's lithography machines. The only one that bought a few was Elcoma, which is the semiconductor and materials division of Philips. A few years later, the founder of ASML, Del Prado, who was deeply in trouble, finally chose to divest, Philips assumed ASML's shares and debts in the joint venture, and cooperated with ASML's bank NMB to accept research and development donations from the Netherlands and the European Community .
Renee Rejimaker, author of "Lithography Giants: The Rise of ASML", commented: "It is absolutely a miracle that ASML is still alive after three years." Obviously, ASML did not fall in the first few years. It is directly related to the blood transfusion. In addition to financial support, Philips also sent its best engineers to assist ASML at the critical moment of developing the Asian market to ensure that the order will not disappear. This is very critical for the start-up ASML.
Of course, Philips is by no means a philanthropist in the field of business investment. In the decades of cooperation, there have been many frictions between ASML and Philips. But it is undeniable that Philips has objectively built a continuously operating shared innovation network. Philips has eyes and ears in every university in the Netherlands, and has contacts with almost every engineering or science professor in the Netherlands. Once the company finds talents in the network of university professors, it will recruit if necessary. When ASML recruited staff at the beginning of its formation, about 300 people applied, which surprised ASML management.
Not only for ASML, but also for the development of high-tech industries in the Wildhofen region has a huge radiation and incubation effect. Dozens of local high-tech companies were directly or indirectly born out of Philips, but Philips did not force them to become subsidiaries, but developed harmoniously with them, and jointly maintained and cared for this shared innovation model. In fact, even the city of Wildhofen has developed based on the growth of Philips.
Even when the well-known TSMC was founded in 1987, it could be regarded as the egg laid by Philips in Taiwan, China. At that time, Taiwan Industrial Technology Research Institute and Philips founded TSMC, and Philips accounted for 27.5% of the shares and was the largest external shareholder. Philips not only unreservedly opened the memory production line to TSMC for learning, but also moved the entire production line to Taiwan to TSMC. It can be said that Philips laid a solid production technology foundation for TSMC, while Zhang Zhongmou mainly created a new business model for TSMC.
Interestingly, because of Philips, ASML received an order for 17 lithography machines from TSMC in 1988, which made ASML barely profitable in 1989. This also proves once again the importance of Philips to early ASML.
Years later, ASML also adhered to the concept of shared innovation favored by Philips. Taking 2016 as an example, ASML allocated 132 million euros to support the construction of its innovation ecosystem, 80% of which was used for R&D work of enterprises, and 20% was used to fund research and development by universities and research institutions.
In addition, ASML cooperates with more than 700 suppliers around the world, 50% of which come from the Netherlands, most of the rest come from the EU and the United States, and 85% of the cost is provided by suppliers. Without the concept of open collaboration, shared symbiosis and its supporting systems, it would be impossible to achieve this. Of course, it is impossible for ASML to gather global wisdom, break through seemingly insurmountable technical problems, and wear the crown of industrial manufacturing.
Join hands with Carl Zeiss to forge a community of shared future in R&D
In 1988, ASML fell into the lowest financial period. At this time, ASML took advantage of the aggressive moves of Nikon and Canon to launch its turnaround - the PAS 5500 lithography machine. Van den Brink, ASML's current CEO, was still chief architect at the time, and he successfully promoted the PAS 5500 to Big Blue IBM, which decided to build a new chip production line in East Fishkill, New York.
After ASML was approved by IBM, orders from all over the world began to pour in. But for ASML, a new problem has emerged, that is, in the face of a steady stream of orders, Carl Zeiss, the core supplier of all projection objectives and lighting systems for itself, has many uncertainties: quality Unqualified, insufficient production capacity, and cooperation with competing products.
Behind the problems of quality and production capacity, one is that the Zeiss management does not really recognize ASML, on the contrary, Zeiss pays more attention to the needs of Nikon; the other is that the Zeiss management is still insisting on manual polishing. Finger" is very proud of it. In this regard, ASML took unqualified shots and used the facts to make Zeiss executives lower their noble heads. At the same time, they also seized the pain point of Zeiss' poor financial situation and promoted Zeiss to transition from "golden finger" polishing to automated operations. .
With the help of new technologies and new equipment such as interferometers, robots, ion beam etching equipment, and polishing pens, Zeiss' century-old production line has been transformed into a flexible production line that meets ASML requirements, and the quality and efficiency of the final delivered lenses have been greatly improved.
After letting the Germans know their mistakes and correct them, ASML also asked Zeiss to supply them exclusively, and the two signed a contract. ASML owns 24.9% of the shares in Zeiss Semiconductor Optics (SMT). Since the fates of both parties are closely linked, when Zeiss CEO Lichtenberg heard the news of ASML’s huge order in 1994, he chose to take a gamble and cut off some business departments. Renovated the production line, and also took a stake in ASML, and borrowed the wind to sweep away the haze of consecutive years of losses in financial reports.
To this day, ASML and Zeiss are still regularly exchanging R&D engineers and scientists to ensure sharing in technology, equity and even corporate culture. Nearly 30% of the total cost of lithography equipment manufactured by ASML was purchased from Carl Zeiss (28.3% in 2018, 26.6% in 2017, and 27.6% in 2016). Zeiss' unparalleled achievements in optical research and development are also helping ASML break through one limit after another.
This is a story of mutual achievement, but without ASML insisting on high standards to force suppliers including Zeiss to continuously upgrade their technology, the pace of innovation in the lithography machine industry would not be so fast; The method ensures the mutual advancement and retreat to the maximum extent, and also proves the importance of financial innovation in cutting-edge manufacturing.
Embrace technology alliances and wear the crown of industrial manufacturing
In 1995, ASML was listed on both the Amsterdam and Nasdaq exchanges. For a while, orders and funds were no longer obstacles to the development of the company. How to defeat the leader, Nikon, became the focus of ASML's management.
In the mid-1990s, if the lithography machine wanted to continue to follow Moore's Law and cross the limit of chip precision, it was necessary to change the light source in the lithography machine from DUV to EUV. In order to challenge this technical problem that was like alien technology at the time, Intel persuaded the Clinton cabinet in 1997 to launch the EUV LLC cooperation organization in the form of a company.
This organization was led by Intel and the U.S. Department of Energy. It not only included the three major national laboratories in the United States, but also brought together Motorola and AMD, which were in full swing at the time. At the same time, Intel also strongly invited ASML and Nikon to join. This is because the old American lithography companies Perkin-Elmer and GCA, which were still struggling in the 1980s, were gone by the 1990s.
However, Intel's move was blocked by the U.S. government because the latter was reluctant to let foreign companies share the most cutting-edge technology in the United States. In this regard, ASML has shown amazing technical foresight, far more aggressively squeezed into EUV LLC than Nikon. ASML lobbied strongly in the United States and offered conditions that the U.S. government could hardly refuse - ASML funded the construction of factories and R&D centers in the United States, and guaranteed that 55% of the raw materials were purchased from the United States, and finally obtained the franchise qualification.
This decision not only ensures that ASML can obtain all the technical patents of EUV LLC, but also ensures that its biggest competitor, Nikon, is completely excluded from the competition. In fact, in the global lithography machine market in 2000, Nikon was still the well-deserved leader, accounting for nearly 70% of the market share, but by 2009 it was surpassed by ASML. Since then, because lithography machines have entered the EUV era, Nikon can only gain a sense of presence in the mid-to-low-end market, and the back of ASML has long been invisible.
From 1997 to 2003, EUV LLC scientists published hundreds of papers in six years, successfully verifying the feasibility of EUV lithography machines, and then announced the dissolution of the alliance. In 2006, the prototype of EUV appeared in the ASML laboratory. Four years later, in 2010, the first human EUV engineering prototype was born in the hands of ASML: NXE 3100.
In 2012, ASML invited Intel, Samsung, and TSMC to invest in R&D, which required 1 billion euros per year, and hoped that everyone would jointly undertake this great human project. In 2015, the mass-produced prototype was released. Although the price is as high as 120 million US dollars, it still receives as many orders as snowflakes. Waiting in line for delivery can take years. In 2018, SMIC had ordered an EUV lithography machine, but the plan had been suspended due to the US technology blockade.
ASML successfully joined EUV LLC in 1997, and attracted customers such as Intel to invest in itself in 2012, which can be said to have brought two major consequences: one is to use EUV to get rid of its competitor Nikon, and the other is to strengthen the bond with the United States, because of political confrontation. Losing China market. For ASML, both of these are unpredictable. It can neither ensure the success of the technical path it chooses, nor can it influence the overall situation of the international political game.
But one thing is certain, that is, if ASML does not join the technology alliance led by Intel and the U.S. government, it will be impossible to obtain the supremacy of the industry; conversely, even if it is as strong as the United States and Germany, today if ASML and TSMC are missing , Samsung, and
It is impossible to realize the closed loop of the chip industry alone.
This, of course, also applies to China.
Some Enlightenments to China's Semiconductor Innovation
For Chinese technology companies, ASML and the Netherlands where it is located have many uniqueness. But in all fairness, this uniqueness does not lie in how many advantages the Netherlands had over China today, but in ASML's ability to maximize its mobilizable advantages.
Today, China has the largest single consumer market in the world. From the government to enterprises, they are also ambitiously trying to conquer the semiconductor highlands. These are all very precious advantages. However, we also need to look at others and deeply reflect on our own shortcomings. As for the case of ASML, it can at least give us the following enlightenment:
First, the outflow of manufacturing supply chains must be reversed as soon as possible. Judging from ASML's experience, the supply chain is its lifeline, and it is also its innovation line. ASML's technical specification and improvement of Zeiss has also occurred in large numbers between foreign-funded enterprises and coastal factories in the past few decades. The most typical example is Apple’s supply chain. In 2008, mainland Chinese companies accounted for only 3.6% of the value of Apple’s supply chain. By 2018, this figure has increased to 25.4%. Behind the data are Chinese supply chain companies collective leap.
Even in the process of integrating into the global supply chain, China has also created its own mobile phone brands. OPPO, vivo, Huawei, Xiaomi, etc. have all enjoyed the multi-faceted spillover effects of Chinese companies' integration into the global supply chain. You know, Philips' earliest lithography machine technology was to follow and imitate American companies, and it took more than 20 years for ASML to defeat Nikon. A sign of its early success was becoming a supplier to IBM.
Therefore, the current exodus of the supply chain of the technology industry needs to be paid enough attention to. It is necessary to pay attention to the growth momentum of Vietnam's import and export data and the number of lighthouse factories in India. These are all obvious signals. Keeping these supply chains, our semiconductor industry has the hope of a breakthrough.
Second, ecological leading companies like Philips should be cherished. Ecological leading enterprises are different from general leading enterprises. They are often the organizers and coordinators of the integration of industry, education and research, and their incubating function for small and medium-sized technology enterprises is incomparable to general leading enterprises and business incubators. For Philips, there is no clear plan to incubate an ASML, but it will naturally build an innovative ecology during its development, and small, medium and micro enterprises bathed in this ecology will naturally get nutrients from the ecology.
In Silicon Valley, the Yangtze River Delta, and the Guangdong-Hong Kong-Macao Greater Bay Area, there are a number of such ecologically-leading companies that build some regional shared innovation ecosystems through equity, knowledge, and talent networks. The reason why Shenzhen and Hangzhou dare to call out the slogan of building an "Oriental Silicon Valley" is that they have cultivated ecological leading enterprises such as Tencent and Ali.
As the economist Fan Gang mentioned in a recent speech, large enterprises are the decisive force for current economic development and industrial progress. There are not too many large enterprises in China, but too few. The research team led by him released the "Shared Innovation Index Report", pointing out that the underlying logic of global economic and industrial organizations is undergoing subversive changes, and technological competition is shifting from competition between enterprises and cities to competition between ecology and urban agglomerations compete. In China, the two deltas have significantly higher levels of innovation than other places, relying on high-density, high-efficiency innovation networks.
Third, we need to realize the importance of the capital market to the semiconductor industry. Europe, where ASML is located, from the Dutch government to the European Community and later the European Union, prefers to subsidize enterprises than the United States. ASML is of course one of the beneficiaries. But ASML really solved the funding problem by listing in Europe and the United States at the same time. After connecting to the global capital market, the company's development quickly entered the fast lane. This point must have been well understood by Chinese companies listed on the stock market between 2010 and 2018.
Relatively speaking, state capital has a longer-term vision, but private capital pays more attention to the positive feedback effect in investment returns. In a rapidly changing market, private capital can better support enterprises in making decisions. After all, no company knows in advance which technology path and which business model can be 100% successful. Therefore, a high-frequency and efficient feedback mechanism is very critical for enterprises, and the signals from the capital market are often the fastest and clearest.
Over the past ten years, from Internet technology to new energy vehicles, a group of Chinese companies have leveraged on the capital market, and have gone through many difficulties and finally made a bloody way to become the darlings of the market. These scientific and technological enterprises with international competitiveness and influence should be regarded as "national strategic enterprises". In the future, China must continue to expand the matrix of strategic enterprises, and semiconductor enterprises will definitely be among the protagonists.
