The global chip shortage, coupled with the rapid rise of Chinese semiconductors, the United States really has to amplify this time.
Last week, U.S. House Speaker Pelosi said the $52 billion chip bill is very close to being ready.
The aim is to reinvigorate the U.S. chip manufacturing industry and ensure that the entire chain of chip production can be controlled.
This news is like a bombshell, stirring up the already restless chip industry even more disordered.
True 10 billion subsidies
In April 2021, U.S. President Joe Biden and global chip industry executives slowly lifted a silicon wafer during a video conference.
He promised the U.S. government $50 billion for semiconductor manufacturing and research.
Of course, the "Sleeping King" is not just talk, accelerating the return of chip production has been put on the agenda, and many key measures have been launched.
On the policy front, the tens of billions of dollars in chip incentive plan is close to finalization.
Last June, the U.S. Senate passed the so-called U.S. Innovation and Competition Act by a vote of 68 to 32 against.
That includes a $52 billion chip grant program to encourage semiconductor giants to build factories in the United States.
Recently, the relevant bill has come very close to being ready, and the specific usage of this $52 billion has been clarified:
$39 billion is used to push chip companies to produce chips and subsidize incentives
$10.5 billion was invested in the implementation of the program
The remaining billions of dollars will be used to build facilities such as chip research and development centers and increase research and development of chip technology
In addition, there is news that the U.S. Congress plans to authorize the CHIPS Act of $190 billion to strengthen U.S. technology and research to compete with China.
If this proposal is passed, the United States will have more confidence.
Echoes
In the face of the "warm invitation" of the United States, the giants of the semiconductor industry are also quite popular.
Up to now, the world's three largest wafer foundries, TSMC, Samsung and GF, have planned to expand their chip factories in the United States, with a total investment of more than $50 billion.
Among them, Intel, which fights locally, is even more wealthy.
With a big hand, he invested $20 billion in advance to build two chip factories in Ohio, USA, creating more than 3,000 jobs.
Construction is expected to begin within this year, with the first fab set to go into production in 2025.
Moreover, this is only the first step, and Intel will invest more than $100 billion in the future to build 8 more chip factories in the surrounding area.
If nothing else, it will become the world's largest chip manufacturing base after completion.
As a key part of the revitalization of the "toothpaste factory", the gold content of this super factory is also quite high.
Although it is not yet known which manufacturing techniques the new fab will use.
But from the process technology roadmap announced by Intel, we can see some clues.
After 3nm, it directly jumped to the next generation process of 20A, leading the chip industry to break through the nano and enter the Amy era (1 Angimon = 1/10nm).
According to Intel's vision, the 20A chip will begin mass production in 2025.
From the point of time, it is not difficult to see that one of the keys to Intel's continuation of Moore's Law is this super factory.
To that end, Intel spent $300 million to book a next-generation lithography machine from ASML to be used to produce chips under 3nm, which are expected to be deployed at the Ohio plant.
Perhaps, at that time, the United States will get its wish and control the world's most advanced chip production line in its own hands.
Obstacles ahead
Of course, the imagination is all beautiful, but there are still many problems ahead of the United States.
First of all, the manufacturing industry of semiconductors in the United States is not developed, and forcibly moving the production line back to the mainland will inevitably lead to an increase in costs.
In an interview, TSMC founder Zhang Zhongmou poured cold water on the United States:
U.S. semiconductors do not have a complete supply chain advantage and will face huge production costs, so the United States cannot succeed in promoting semiconductor localization.
This is true, turning over the data, you will find that the United States' share of the global semiconductor manufacturing industry has fallen from 37% in 1990 to 12% today.
Moreover, Zhang Zhongmou also has confidence in saying this.
At present, TSMC has a market share of 55%, accounting for half of the global chip foundry field.
The second-ranked Samsung has an 18% share, and the two Asian companies combined monopolize 73% of the chip foundry market.
Not only that, 5nm and more advanced chips, only these two giants can produce at present, it is no wonder that the United States is in a hurry.
Semiconductor industry player Zhi Guoping said to the daily push of technology that Intel wants to surpass TSMC, which is very difficult.
One is because of Intel's model, and the other is because the learning curve is not enough.
Zhang Zhongmou analyzed in a speech that the learning curve is TSMC's most important advantage.
The so-called learning curve is that with the increase of the cumulative output of products, the cost of unit products will decrease in a certain proportion, which is simply that practice makes perfect.
Due to the lower production costs in Asia and the closer proximity to downstream customers, it has naturally become a base camp for chip foundries.
TSMC takes advantage of its scale and continuously reduces prices through the learning curve to gradually form its own competitiveness.
The semiconductor companies in the United States have evolved into "fabless" chip companies, which are only responsible for the design part.
Chip production has long relied on TSMC and other generations of enterprises, resulting in the shrinking of the local chip manufacturing industry, and the production-side talent training mechanism has also been broken.
The most typical example is Intel, which has strong EDA development and chip design capabilities.
But in the face of the learning curve, "banknote ability" sometimes becomes "banknote power".
Due to Intel's lack of experience in manufacturing cutting-edge chips and the lack of learning curve, even if it has money, it is difficult to catch up with TSMC in the field of chip manufacturing.
What's more, Intel's products are mainly PCs, servers and data centers, which are not the forefront of chip demand and cannot promote the production of advanced chips in the United States.
In addition, the United States also faces competition from other countries and regions.
As a strategic technology in the future, semiconductors are extremely competitive in the world, and it is not only the United States that covets the leading position of chip manufacturing, but also South Korea, Japan, and the European Union.
The European Union plans to produce 20% of the world's chips by 2030.
South Korea is even more ambitious, proposing the "K Semiconductor Strategy".
In the next decade, it plans to invest 450 billion US dollars with Korean companies such as Samsung Electronics and SK Hynix to build South Korea into the world's largest semiconductor production base.
$450 billion, equivalent to a quarter of South Korea's GDP last year, compared to $52 billion in the United States.
Will the United States be left behind in the competition?
After all, in the US "policy toolbox", there are more ways.
Last year, on the pretext of improving the transparency of the chip's "supply chain", the United States took a strong attitude and demanded that TSMC, Samsung and other wafer foundries hand over confidential data such as technical nodes, yield rates, inventory quantities, and sales records.
To urge companies to "voluntarily" hand over information, the U.S. Secretary of Commerce said:
If businesses are reluctant to submit, there are other ways in our toolbox to get them to give us the data. I hope we don't get to that point, but we will act if necessary. ”
Later things, we all know.
Under pressure, more than 20 companies such as TSMC, Magnesium, and Samsung finally chose to submit relevant materials.
Once the United States has this data, it can clearly understand the future chip layout of competitors and formulate countermeasures.
It is equivalent to opening a perspective plug-in and wantonly breaking the rules of the game.
Pressure came to China's side
The United States intends to occupy the commanding heights of chip technology through a soft and hard approach, which is not good news for China.
Data show that in 2020, China's semiconductor imports will be 350 billion US dollars, and the self-sufficiency rate will be only 16%.
If you exclude TSMC, Samsung, Hynix and other companies that have set up factories in Chinese mainland to produce chips, the real self-sufficiency rate is probably only 10%.
Although China has a certain advantage above 28nm, it is still difficult to get rid of its dependence on imported semiconductors.
Huawei, whose shipments have fallen to Others due to sanctions, is a living example.
Recently, Magnum dissolved the Shanghai R&D Center and wanted to immigrate to the United States with its core Chinese employees.
Of course, we didn't sit still.
According to the daily push of industry insiders to science and technology, China's investment in semiconductors is far greater than that of the United States.
As early as 2014, the Ministry of Finance led the establishment of the National Integrated Circuit Industry Fund, with an investment of 100 billion yuan in the first phase and 150 billion yuan in the second phase.
It can be seen on domestic chip leading enterprises such as SMIC and Huiding.
Cartography @ Billion Euro
Interestingly, Moonbird noticed that behind this big fund, there are also a bunch of national brand enterprises contributing.
One of the most conspicuous is the China National Tobacco Corporation.
—Finally—
"There is only one true heroism in the world, and that is to recognize the truth of life and still bravely love it."
--Romain Rolland
Perhaps there is no industry that can burn more money than semiconductors.
If you want to build a semiconductor company that can gain a firm foothold in the world, you must have the consciousness of not making money for at least ten years, and you must burn a lot of money into it.
This is true of Samsung, and it is even more so of TSMC.
Of course, in addition to burning money, we also have to learn the advanced methods of others.
China's semiconductors still have a long way to go, and the United States will not relax its suppression of China's high-tech.
However, isn't the road to climbing upwards more exciting than standing at the top?
END
Source丨 Technology daily push
Welcome to click on the video number of Easy Jane Finance to watch the latest video~