The author of this article is a guest of the studio [Captain Xinghai]
one
In 1979, Acorn, a Cambridge-based start-up computer company, received an unexpected job.
A customer approached them and wanted them to produce a slotmachine, and everyone who had seen the old Hong Kong film heard the name and understood that it was a slot machine.
Soon, a slot machine based on the 6502 CPU (which is the famous red and white CPU) came out, and Acorn got the first pot of gold.
Since the performance of this slot machine is very good, Acorn had a whim, and put this slot machine with a keyboard and turned it into a computer, can it wrestle with Apple's Apple II?
Just do it,So the BBC Micro computer modified from the slot machine was born,This thing is quite simple,The basic version doesn't even have a monitor,Just a host integrated with the keyboard,Connected to the TV to realize some basic functions of the computer,To be honest,It's basically the same as the little overlord learning machine。
Although the BBC Micro was successful, it was a dick compared to the Apple Lisa, which pioneered all-in-one computers at the same time, and the profit margin was pitifully low.
So, Acorn wanted to develop an all-in-one computer like Apple, but soon found that the 6502 CPU could not meet the needs, so Acorn approached Intel to see if the Intel 80286 CPU architecture could be used in their new computer?
What Acorn wants to buy is not a CPU, but a CPU architecture, and then under this architecture, it uses Intel's design and instruction set to deeply customize it according to its own needs.
This kind of authorization doesn't make much money, Intel didn't even look at it and said,.
History has shown that Intel made a very wrong decision and single-handedly cultivated a formidable opponent.
Acorn, who had eaten the closed door, came back and learned from the pain and made up his mind that I would develop my own CPU!
Soon, Acorn got its own chip, that is, Acorn RISC Machine, referred to as ARM1, yes, this is the later famous ARM, the performance of the chip is not bad, but more importantly, Acorn has made a set of reduced instruction sets that are completely different from Intel x86-RISC.
The concept of instruction set is complex, and we can simply understand it as the operating specification of a machine's hardware. For example, if an army is a computer, then all soldiers are hardware, soldiers' various tactical codes and rules and regulations are instruction sets, and the army's tactical plans are software.
When the software starts running, the tactical plan should be broken down by various tactical drills and rules and regulations, guiding each soldier to do what he wants to do, and when each soldier completes his or her own decomposed task, the combat mission is completed, and in this process, the soldier does not even need to know what the tactical plan is, he only needs to poke forward with a spear.
Why do we have an instruction set? This is to ensure that the same software can run on different hardware.
If the set of orders is different, it is like handing over the tactical plan of the Roman army to the Qin army to execute, which will definitely not be implemented, but if the Qin army and the Roman army both use the same set of "tactical rules and regulations", then it will be no problem for them to implement the same tactical plan.
Later, Intel and ARM went their separate ways, and Intel, with its de facto alliance with Microsoft (the Wintel ecosystem), made its x86 instruction set a mountain in the computer world.
Although ARM's RISC debuted late, it has formed an AA ecosystem with the Android operating system by relying on its own open-source characteristics, and occupies the vast majority of the mobile device processor market.
In the past few decades, the x86 instruction set and ARM's RISC instruction set have occupied a monopoly position, and no matter who wants to make chips, they can't avoid these two, even Apple's M series high-end chips are also RISC.
China's high-end processor chip companies are no exception, either based on the existing architecture of overseas companies for research and development, or through intellectual property (IP) licensing to design chips, mainly ARM.
In this way, it is equivalent to giving the neck to someone else's hands, so that others can pinch it as much as they want.
In 2021, Nvidia acquired ARM.
Although Nvidia said that the acquisition will not affect the ARM authorization of Chinese chip companies, looking at the degree of the current US sanctions on Chinese chips, who can guarantee that one day in the future, Nvidia will not cancel the IP authorization of Chinese chip companies?
Therefore, Chinese must have their own instruction set, such as Huaxia Core (Beijing) General Processor Technology Co., Ltd., founded by Li Keyi, which once claimed to be the only company in China with a self-developed instruction set.
It stands to reason that such a company should have a bright future. The truth is, however, that it is bankrupt. But it is not the only one that has gone bankrupt, in 2023, 10,900 chip-related companies in China will disappear, an average of nearly 30 per day, which can be called a wave of chip company closures.
What's going on with all this?
two
There are three key nodes in the development of China's chip industry in recent years:
The establishment of a large fund, the establishment of the Science and Technology Innovation Board, and the sanctions imposed by the United States.
In September 2014, the National Integrated Circuit Industry Investment Fund was formally established, although the fund is completely market-oriented, but its camp is quite luxurious, the largest shareholder is the Ministry of Finance, and other shareholders include many well-known state-owned enterprises and central enterprises, such as China National Tobacco Corporation, China Mobile, etc.
Because of the huge scale involved (the first phase of fundraising of 138.7 billion yuan), it is called a "big fund" by the industry.
The operation direction of the large fund is to improve the construction of the supply chain supporting system of the integrated circuit industry, and invest in the whole industrial chain such as chip design, wafer manufacturing, packaging and testing, special equipment and core components, key materials, and ecosystems.
Translated into these words, it is to bring together some wealthy domestic enterprises to become the big financiers of China's chip industry, alleviate the bottleneck of investment and financing in the development of China's chip industry, and enhance the confidence in the development of the industry.
Judging from the timing of the establishment of the big fund, it is not difficult to find that China's decision-makers have a very accurate judgment of the international situation, and if they act a few years later, the situation will be very different.
The biggest role of the big fund is to lead a part of the social capital to turn from real estate and the Internet to the chip field, which has triggered the entrepreneurial upsurge of domestic chip enterprises.
In 2019, the birth of the Science and Technology Innovation Board activated the chip entrepreneurship fever, and even a small company could easily raise funds on the Science and Technology Innovation Board, so chip entrepreneurship began to blossom everywhere.
On May 16, 2019, the U.S. Department of Commerce put 70 companies on the "Entity List" on national security grounds, prohibiting U.S. companies from selling related technologies and products, opening the curtain of chip sanctions.
The huge market gap after the sanctions has once again stimulated more social funds to invest in the chip industry, and the investment in China's semiconductor industry has soared from about 30 billion yuan in 2019 to 387.6 billion yuan in 2021, an increase of more than ten times.
If the upstream and downstream industries are counted, the big fund uses more than 130 billion yuan to leverage the market of up to one trillion yuan.
One of the things that was once pinned on high hopes was Ziguang.
Zhao Weiguo, the former boss of Unigroup, graduated from the Radio Department of Tsinghua University (later the Department of Electronic Engineering), and most of his brothers and sisters are the mainstays of China's chip industry, such as Yu Renrong of Weir Co., Ltd., Shu Qingming of GigaDevice Innovation, Feng Chenhui of Zhuosheng Microelectronics, etc.
However, compared with them, Zhao Weiguo's development path is different, he first participated in the establishment of Tongfang Microelectronics in Unigroup, but it didn't take long for him to jump ship to Jiankun Investment and change his career to real estate.
Fortunately, Zhao Weiguo caught up with China's real estate boom, and the investment was a huge success, in his own words, "At that time, entering real estate was like grabbing money, I took 1 million yuan to Xinjiang, and when I came back, I had already earned 4.5 billion yuan, 4,500 times!"
Later, in 2009, the poorly managed Unigroup implemented a mixed-ownership transformation, and Zhao Weiguo relied on the injection of capital into Unigroup to get the management power of Unigroup and became the chairman.
Since then, Zhao Weiguo has transformed from a real estate tycoon to a "chip maniac" who amazes everyone, and he has threatened to spend $23 billion to buy Micron, $3.8 billion to buy a 15% stake in Western Digital, to buy MediaTek, and even to take a stake in TSMC.
Although Zhao Weiguo was ridiculed by Guo Taiming as "a stock speculator", he did spend hundreds of billions of dollars to acquire 16 chip companies, making the scale of Unisplendour swell from 1.3 billion yuan in 2009 to 296.6 billion yuan in 2020.
That is, in 2020, Zhao Weiguo launched a new generation of high-performance security chip THD89 to fill the domestic gap.
For a time, Zhao Weiguo became the light of the nation and was known as the second Ren Zhengfei.
At this time, Unisplendour's business scope covers mobile phone chips, memory chips, programmable chips, FPGAs, integrated circuit packaging and testing, etc., and is considered by the industry to be the main force in the rise of China's chip industry.
So the question is, Zhao Weiguo buys chip companies everywhere, where does the money come from?
The answer is simple, big funds.
Unisplendour Communication, a wholly-owned subsidiary of Tsinghua Unisplendour Group, is one of the initiators of the big fund, and has an innate advantage when taking investment. But the problem is that if you only spend money on mergers and acquisitions, you can't absorb the technology brought by the acquired enterprises, how can the company make a profit?
Obviously, Unigroup has not done a good job in the marketization of absorption technology.
In 2020, Tsinghua Unigroup exploded, and in July 2021, under the application of creditors, Tsinghua Unigroup entered the bankruptcy reorganization procedure, triggering a series of avalanche effects, which many people call "photo curse".
On September 11, 2018, Ding Wenwu, President of the Big Fund, and Lu Jun, President of Huaxin Investment (Big Fund Manager), went to Unisplendour for investigation, and Ding Wenwu and Lu Jun said that the Big Fund and Huaxin Investment will continue to support the development of Unisplendour and continue to deepen cooperation in the future.
After the research, they took a photo in front of the background wall of "Changing the World with Dreams". From left to right, the group photo is Diao Shijing, Co-President of Tsinghua Unigroup, Lu Jun, President of Huaxin Investment, Zhao Weiguo, President of Tsinghua Unigroup, Ding Wenwu, President of Big Fund, Zhang Yadong, Vice President of Tsinghua Tsinghua Unigroup, and Qi Lian, Co-President of Tsinghua Tsinghua Unigroup.
These 6 people were known as the Chinese chip group at that time.
No one expected that only 4 years later, 4 of the 6 people in the photo completely overturned.
Lu Jun, Diao Shijing, Zhao Weiguo, and Ding Wenwu were taken away for investigation......
At present, the results of the investigation of Ding Wenwu and Diao Shijing have not been released, but judging from the official charges of Lu Jun and Zhao Weiguo, it is not difficult for us to analyze some clues.
Lu Jun's mistakes in involving business were: deviating from his main responsibilities and main business, resisting organizational review; violating integrity and discipline, illegally investing in a loan client company, and using his power to seek benefits for the business activities of his relatives and friends; alienating the power in his hands into a tool for seeking personal gains, taking advantage of his position to seek benefits for others, illegally accepting property, misappropriating funds, abusing power, and acting recklessly, causing serious losses to the state.
Zhao Weiguo's mistake was: As a manager of a state-owned enterprise, he was consumed by the desire for profit, acted recklessly, betrayed his duties and missions, used public weapons for private purposes, turned public affairs into private interests, regarded the state-owned enterprises under his management as private territory, deliberately seized state-owned assets, handed over the profitable business of his unit to his relatives and friends in violation of regulations, purchased commodities from the units operated and managed by his relatives and friends at prices significantly higher than the market price, and instructed the directors of the listed companies to carry out acts that harmed the interests of the listed companies, causing particularly heavy losses to the interests of the state.
See that both of these qualitatives have brought losses to the country, which is obviously related to Ziguang. The state wants them to concentrate on big things, but they use the national strategy as a tool for personal gain.
If it is only these tens of billions of losses, the impact may not be large, but the key is that their greed has invested too much money in companies that do not have relevant capabilities, and those who have the ability cannot get investment, delaying the precious time for the rise of China's chip industry.
From this point of view, they are not to blame for their deaths.
叁
Of course, there is no need for us to be harsh on the big funds, because the big fund case is the fault of people, not the big funds.
Although this matter has made many people point the finger at the big fund, we must also see that it has provided financing to a large number of domestic chip companies, Changdian Technology, North Huachuang, China Micro Semiconductor and other well-known chip companies have received large fund investment, some major breakthroughs in the domestic semiconductor industry in recent years, in fact, there are large funds behind it, and most of its investments have achieved investment returns, realizing the promotion of the development of the chip industry and the preservation and appreciation of state-owned capital.
Therefore, the national level did not beat the big fund to death, but insisted on starting the follow-up project of the big fund.
However, compared with the corruption in the implementation of large fund projects, the impetuosity of the domestic semiconductor industry is more serious and needs the most attention.
After the establishment of the Science and Technology Innovation Board, the venture capital of domestic semiconductor companies has a relatively clear exit mechanism, attracting a large amount of social capital investment. The U.S. sanctions and domestic substitution strategy have made the market demand large and strong, and have also attracted a lot of social capital to invest in chip companies.
As a result, chip companies began to bloom everywhere, and if we check it ourselves, we will find that many of these companies are cross-border to make chips, such as cross-border real estate industry, cross-border financial industry, and even some pig breeding companies have followed the trend to engage in chips.
Tianyancha data shows that there will be more than 495,000 chip-related enterprises in China in 2023, of which more than 112,000 will be newly registered in 2022, and the growth rate of new enterprise registration will reach 32.7%. Do you think there are many companies with feelings for the country and the ideal of serving the country with science and technology, or are there many companies that want to take advantage of the wind to make quick money?
There is no need to blame for following the trend to make quick money, but the problem is that the particularity of the chip industry is inherently contradictory to the pursuit of profits by capital.
The chip industry has a high technology content, which not only requires a high-level team, but also needs to continuously develop new technologies, which requires a large amount of capital investment. In addition, there is a need for continuous investment in updating equipment, improving processes, and developing new products. It generally takes several years or even longer for a product to go from R&D to production to promotion.
Therefore, high investment, continuous investment and long return cycle determine that if you want to make quick money from chips, you have to take a lot of risks.
Let's just give one example to understand, tape-out.
In layman's terms, tape-out is trial production, the chip company will hand over the designed plan to the foundry, produce some samples to test, see if there is any problem with the design plan, and if the test passes, it will start large-scale production in this way.
But the problem is that tape-out is too expensive.
The tape-out process is very complex, and the process, wafer, and chip design used may affect the success rate of tape-out.
It has been estimated in the industry that 14nm process chips require about $3 million for tape-out, 7nm process chips require $30 million for tape-out, and 5nm process chips need $47.25 million for tape-out.
Therefore, for some start-up chip companies, tape-out is a huge risk, and if it fails once, the entire company may collapse.
In order to reduce risks, chip companies with capital investment must be responsible for shareholders and take a short-term route. To put it simply, it's the easiest to pick.
As we all know, chips generally consist of three main links: design, process, packaging and testing.
The design is mainly the circuit architecture design of the chip, including lithography, etching, deposition, cleaning, etc., and packaging and testing is the packaging, testing and heat dissipation of the manufactured chip.
Of these three links, manufacturing is the most difficult, because the lithography machine is stuck in the neck. Packaging and testing is second, but in the field of packaging and testing, the gap between China and the world's top companies is not large, and the simplest and least invested is design.
Let's put it this way, even a college student, as long as he has an outsourcing team in charge of technology and holds a few open source design drawings, as long as he can operate successfully, he can raise funds in the chip tide and obtain hundreds of millions of valuations.
In 2021, there will be 135 chip companies listed in China, including 8 packaging and testing companies, 4 manufacturing companies, and 82 design companies!
In fact, these 82 companies are still a little strong, at least they are listed, and more chip design companies behind them are completely indiscriminate, and there are not many that can complete tape-out, let alone produce good chips. When the company's development momentum changed from technology-driven to financial-driven, no one mentioned Xingchen Haihai, and only pursued practical interests.
As a result, China's much-needed chip manufacturing field is unattended, and the chip design field is in full swing; high-end chips are in short supply, and low-end chips are in surplus, and all kinds of disorderly competition in turn worsens the market, leaving a large number of companies that have been doing well in the past into losses.
What's even more excessive is that the original intention of the establishment of individual companies is not to make chips at all, but to "cheat subsidies" and "cheat financing".
For example, in the famous Hongxin scam, Chairman Long Wei is a college student, and the director Li Xueyan turned out to be selling soju, opening restaurants, and reselling traditional Chinese medicine, and the director Cao Shan, who only has a primary school diploma. The funniest thing is Supervisor Li Yueru, who turned out to be Li Xueyan's "personal nanny".
Just such a pheasant team has to invest 128 billion yuan in chip manufacturing projects at 14nm, 7nm and smaller nodes, threatening to become the most advanced chip manufacturer in China.
Because the story is well told, this project has become a "provincial key construction project", and has obtained a lot of land and preferential policies. As a result, in 2021, Hongxin exploded, and the 100 billion scam was wiped out, leaving only a bunch of unfinished buildings. The lithography machine was useless for a day after it was bought, and after a grand entrance ceremony, it directly became collateral.
In addition, there are Nanjing Deke Ma, Chengdu GF, Guizhou Huaxintong, Shaanxi Kuntong, and Jiangsu Dehuai, all of which ended up in one place.
If we look at the 10,000 chip companies that will collapse in 2023, we will find that the vast majority of them are new companies that have been established for a year or two, that is, the kind of companies that "make quick money".
These situations are actually very normal, after all, the chip industry is not the Internet, the wind of the Internet can blow pigs, but it can't blow the chip industry, whether it is a big fund or private capital, it can't stand the continuous burning of money but can't see the output for a long time, maybe it can barely support in the face of the national strategy, but once it encounters the fluctuations of the world semiconductor market, it will die a large piece.
In 2020 and 2021, when the new crown epidemic spread, home office stimulated the huge demand for consumer electronics products such as mobile phones, computers and TVs, which led to a sharp increase in chip shipments.
At the same time, the Federal Reserve's interest rate hike and high inflation have also weakened residents' ability to spend money, and also suppressed the demand for chip products on the consumer side.
So in 2022 and 2023, the world chip market will usher in a cold winter, and according to the World Semiconductor Trade Statistics Organization (WSTS), the global semiconductor market size is expected to decrease by 9.4% year-on-year to $520.126 billion in 2023.
There are no eggs under the nest. The industry has turned from prosperity to decline, even Samsung's profits have to fall by ninety, and even TSMC has to face cutting orders, not to mention those domestic start-up chip companies that are operating dismally? In this way, continuous losses will inevitably consume the enthusiasm of capital, and there are not a few people behind the financiers who withdraw their capital and run away, and a large number of chip companies have died.
The collapse of 10,000 chip companies in 2023 is the inevitable result of this general situation.
肆
After reading so many deplorable stories, many people will wonder, is there any hope for Chinese chips if it goes on like this?
In fact, there is no need to be discouraged, survival of the fittest is a natural law, and the closure of some poorly managed chip companies can reduce disorderly involution competition in the industry, and objectively increase production capacity, which is conducive to companies that are truly valuable and serious about making chips.
More importantly, the determination at the national level has not changed, and it has not given up on the big fund because of the problem of the big fund, but has increased investment, according to Reuters, the third phase of the big fund plans to raise 40 billion US dollars, about 300 billion yuan.
From the perspective of the investment direction of the previous large funds, the first phase focused on the field of chip manufacturing, focusing on memory and advanced process production lines, and the second phase focused on the new product industry chain, while increasing the proportion of investment in the field of equipment and materials, covering EDA, wafer manufacturing, packaging and testing, chip manufacturing equipment, materials and applications and other industrial chain links.
It can be said that these investments have borne fruit.
In the field of chip design, Kirin's return has proven the progress of China's chip design, and more importantly, the domestic mobile phone industry, which consumes large chips, such as Xiaomi, OPPO, and VIVO, are developing their own ISP, NPU, AI and other chips. Many other companies are also vigorously developing RISC-V chips, and they also want to get rid of the dependence on ARM, x86 and other instruction sets as soon as possible.
In the field of subdivided consumer electronics chips, many chips have gotten rid of foreign dependence, and have even begun to counterattack foreign markets, such as Weir's image sensor chips, GigaDevice's NOR flash memory chips, Shengbang Microelectronics' analog chips, Cambrian's AI chips, Goodix's touch chips, Hengxuan Technology's Bluetooth headset audio SoC chips, etc., have all caught up with the world level.
In the field of chip manufacturing, the yield rate of 14nm is stable, 7nm has been put into production, and higher-process chips have also achieved breakthroughs, as evidenced by the Kirin 9010 chip on the Pura70.
The gap between the chip packaging and testing field and the outside is the smallest, and JCET has become the world's third largest packaging and testing manufacturer, and it can catch up with the leader in the short term.
Of course, the biggest gap at present is in the chip equipment and materials industry. The chip manufacturing equipment produced in China is still being debugged at the 28nm stage, and the yield rate is not as good as that of ASML. But to be honest, many chips are not as practical as the high process, and 28nm is enough.
What is really stuck is that some chip materials such as photoresist, etc., cannot get rid of foreign dependence.
This is actually a dilemma for China's large funds to invest in chips, China's chip industry as a whole is relatively weak, whether it is design, equipment, materials, manufacturing needs investment.
However, the chip industry is like a wooden barrel, you can't shorten any board, and the crazy sanctions of the United States have forced the direction of large fund investment to be comprehensive. This leads to the fact that although the big fund has a lot of money, it is still not enough to spread it evenly over the entire industrial chain, which is also something that cannot be helped.
Moreover, the investment of large funds, to put it bluntly, is a government act, since it is a government act, then all the shortcomings of government behavior (such as bureaucracy, blind investment, performance projects, etc.) will be reflected in the investment.
As a result, large fund investments are in a state of flooding, and it is difficult to ensure that every investment can be accurate. We can only hope that the involution after the flood irrigation will allow the real strength to emerge, and it is difficult to avoid deception and duplicate construction in this process.
However, it doesn't matter if there are some cheats and duplicate construction, as long as someone can stand out, this is actually the same idea as the "saturation rescue" in the wandering earth, no matter how much the cost is in the process, as long as the final effect is achieved, it is a victory.
Let's look back at the photovoltaic, new energy vehicles and LCD panel industries, which has not experienced subsidies, savage growth, and bankruptcy, and finally ushered in the rise?
Of course, some problems in the middle, such as corruption, such as fraud and compensation, must still be supervised, otherwise bad money will drive out good money, and no one will work seriously.
Therefore, we do not need to worry about the collapse of 10,000 Chinese chip companies, because in the same year of 2023, the number of newly registered chip-related companies in China reached 65,700.
10,000 people fell, 60,000 stood up, it's not a big deal, after the survival of the fittest, what remains will naturally be high-quality.
What's more, because of the sanctions imposed by the United States, the domestic chip market demand has generally carried out domestic substitution, which is equivalent to the United States giving up a trillion-dollar chip market to Chinese chip companies. At the same time, the rise of AI technology has also created a huge demand for computing chips, and only China and the United States have the ability to catch up with the AI outlet.
We believe that in the face of such broad market prospects, coupled with national industrial support, Chinese chip companies will be reborn and usher in a huge outbreak and breakthrough.
And we just need to wait for the flowers to bloom.