Intel, which has just announced that it has skipped the 10-nanometer process and will postpone production to 2019, was shocked again. Recently, the industry has continued to report that in the past few months, the Intel foundry department has taken orders to brake the brakes. This has puzzled the industry: "Is Intel not wanting to play the foundry?"
In fact, this is not the first time Intel has reduced or faded foundry business. In the past, Intel had a similar precedent. When the demand for own product shipment capacity was tight, it was a strategic reduction in wafer foundry. Business orders. But it is worth noting that this time is obviously different from the past.
Because, in the past few months, Intel's focus has clearly shifted to the production of its own internal products, and continued to reduce foreign orders for OEM orders. There are many signs that Intel may gradually fade out of the global foundry business. This is undoubtedly a super blockbuster for the global industry.
And if Intel really fades out, or even withdraws from the foundry business, many people may be the first to think of benefiting from TSMC, but in the view of DT, Samsung may be the biggest beneficiary. The huge turning point is absolutely the focus that domestic semiconductor companies must concentrate on.
If Intel really gradually fades out of the OEM market, the original situation of TSMC, Intel, and Samsung will evolve into a match for the heroes. Domestic companies such as SMIC and Hua Hong Semiconductor must strengthen their own strength and catch up quickly in order to be able to compete in this field. In the industrial sector, it will take advantage of the opportunity to fill up the vacancy.
Processor security vulnerabilities, multiple batches of tickets on 10nm process, Intel frequently encounters bottlenecks
Intel is in an uphill battle. In addition to the outbreak of processor security vulnerabilities, the progress of the latest generation of 10-nanometer process has been repeatedly delayed. It was four years ago that the previous generation of 14-nanometer production was launched. Compared to its major competitors, TSMC and Samsung It has already entered the 7-nanometer melee combat. This is an extremely unusual bottleneck for Intel, which has been arrogant to global semiconductor hegemony and has been at least three years ahead of its competitors.
Three years ago, Intel beat the drums to return to the foundry industry. At one point, TSMC had compared Intel and Samsung's two rivals to the "hundred pound gorilla". However, all this has changed. Since early this year, Intel has significantly reduced the number of external foundry orders, but is this a temporary fade-out or a permanent decision? There is no answer at the moment.
Intel, which has been in the semi-conductive world for half a century, has spent 10 years to enter the foundry market. In the end, it may only be a dream. DT Jun thinks that there are two major factors worthy of attention.
The proximate cause is the delay of process technology. Intel only announced 10 nanometers until 2019. It was four years ago that the previous generation of 14 nanometers was launched. The slowest pace in this industry is that it takes an average of two years to advance a process generation. Intel’s technology is too late. exaggeration.
Although Intel has announced that its 14nm is equivalent to the competitor's 10nm, TSMC's 16nm is actually 19nm in terms of transistor linewidth, and 7nm is equivalent to Intel's 13nm. However, today Samsung and TSMC are already To mass-produce 7 nanometers, and in the past, Intel was proud of the technology's three-year distance from its rivals. This time, Intel's technical bottlenecks are getting harder and harder to explain.
Intel's 10-nanometer yield cannot be solved, resulting in a serious shortage of existing 14-nanometer and 22-nanometer production capacity. It is unable to meet the needs of external customers. Even the production capacity is insufficient. How can we help OEM customers?
Figure 丨 Intel 10nm process bottleneck
Furthermore, Intel recently broke a major processor security vulnerability. The US Computer Network Crisis Management Center even recommends that users should replace the affected processor. This may cause a wave of replacement surges in the coming years. In the industry processor business, light production of their own products may not be completed, and there is still a sense of leisure to meet OEM orders.
A culture of lack of service awareness, destined to Intel's foundry business ten years empty
The reason why Intel’s foundry industry could not succeed was that the chairman of Taiwan Semiconductor Manufacturing Co., Ltd. Zhang Zhongmou had spoken in a word. He analyzed that Intel's own product (processor) is very powerful, but the foundry is to help others make products, we must have the spirit and culture of serving others, but Intel will only serve itself, will not serve others.
According to industry analysts, Intel has always dominated the semiconductor industry. Regardless of technology and products, it swept away thousands of troops in the computer era, did not push AMD into a desperate situation, and just wanted to get rid of some anti-trust concerns. Basically it is an invincible hand in the world, so the mentality has always been very proud.
When selling their own products, "confidence" is a plus, but to do OEM must often be bent down, expect others to use their own foundry capacity, technology, to serve others and meet customer needs, which Point is Intel's biggest weakness.
In fact, those who lead Intel's foundry business are all from old Intel. They have inherent arrogance in their bones. They just say that they really don't understand the true meaning of the foundry. It is clearer that Intel is There is simply no right person to dominate the foundry business, so it will not succeed.
Intel’s self-confident strong players will naturally not be willing to lend to foreigners. Therefore, it has always been heard that TSMC will reuse Intel’s talents to learn its expertise, but with Intel’s superior attitude, it will not take the initiative to dig TSMC. People learn the secrets of foundry business.
However, perhaps Intel has really encountered too many challenges recently. Recently, it began to hunt for talents from other big manufacturers and found Jim Keller, who once worked for AMD and switched to Tesla in 2016. The industry insiders think this is a good start. In the future, Intel will have the opportunity to use a large number of external talents to further exchange blood for the company and accelerate its metabolism.
Fig. Jim Keller
In fact, Intel opened its foundry business in 2008. But for ten years, it has been operating the business as a sideline, or as a buffer for the adjustment of its own processor's main business. For Intel, the foundry may be It is strategically important, but it has not always been considered as the main core, but perhaps it is precisely because of this mentality difference that the Intel foundry business is not doing well. Until 3 years ago, Intel announced again that it must seriously enter the foundry field.
When Intel and Spreadtrum teamed up for two 14-nanometers that made a stir, did LG suffer?
What are the founding customers of Intel? Their biggest ally is Ziguang Group's Spreadtrum and South Korean mobile phone maker LG.
In the past two years, Intel has successively cooperated with Spreadtrum on two 14-nm manufacturing products. This alliance is considered as an important battle for Intel to enter the field of Asian foundry and is also the touchstone for Spreadtrum to enter into high-end products. It is also because Intel has already invested in the exhibition information before, and the intention of forming an alliance and assisting in gang fighting is quite obvious.
If you mention the victims of Intel 10nm repeated bounce tickets, you have to mention LG. LG's 10-nanometer mobile phone processor finally chose to cooperate with Intel. Not only did it invest a lot of money, but it took two years. However, because Intel's 10nm chip has been bounced, the last LG fund had been burned and they had to give up the plan and let LG make it. The processor department is seriously injured.
In addition, Intel’s foundry customers include Achronix, an FPGA vendor, and FPGA maker Altera, which has been acquired by Intel.
Although Intel has significantly reduced the number of orders received from foundry customers, it still continues to undertake its own internal and group products, such as baseband chips. However, this strategy has indirectly affected TSMC.
Although Intel and TSMC are rivals in the foundry industry, Intel is also one of TSMC’s major customers. Intel’s own baseband chip is produced by TSMC. This is a very delicate relationship. Intel’s own technology, There is capacity, but some products still have to be handed over to competitors.
If Intel fades out of the foundry territory, the long-term biggest beneficiary is Samsung rather than TSMC.
If Intel withdraws from the foundry industry, the beneficiaries in the short term will be TSMC and Samsung. From the top three killings to the two confrontations, in the long run, this situation may cause Samsung to have a bigger chance in the foundry industry. , and further strengthen the strength of the global semiconductor field, is not conducive to TSMC.
However, the most miserable is definitely a lot of chip design companies. In the future, there will be fewer and fewer suppliers with high-end manufacturing processes, and the bargaining power is getting lower and lower. What's more painful is that many chip design companies may have to be forced to Go to Samsung to cooperate.
In addition to the major foundries, the relatively small-scale chip design companies will certainly look for the second supplier. The main foundries are usually TSMC. In the past, the role of the second supplier was UMC, but UMC has already failed to fight high-end products. Process, so in recent years by Samsung, Intel, GlobalFoundries fill in the second supplier role.
If you think about it, the only high-end process suppliers in the world other than TSMC will be available only from Samsung, Intel, and GlobalFoundries. Now that Intel does not play, GlobalFoundries often changes its policy to make customers worry that the only technology is that Samsung, the only manufacturer that has production capacity and manufacturing process, may eventually be forced to go to South Korea for filming. This is tantamount to making many chip companies fall into a dilemma.
Figure 丨GlobalFoundries
Furthermore, Intel, GlobalFoundries, the two semiconductor giants, do not want to take the lead in the mentality of operating foundry. For Intel, the main business is still its own processor products. The foundry business is picking customers and only looking for specific strategic partners. The reason why GlobalFoundries does not want to grab the leader is actually because it doesn’t matter. Because of the difference in innate strength.
However, Samsung is not the same. Samsung is the first company to do whatever it takes. No matter if memory, flash memory, panels, mobile phones, or TVs are all the world’s top goals, who would believe that Samsung is making waves on its wafers? The foundry industry will be willing to live behind TSMC.
However, when Intel also faded out of the foundry industry, it was tantamount to many customers losing a high-end technology supplier. Finally, they were forced to go to Samsung to do OEM. Therefore, this situation will benefit TSMC and Samsung in the short term. In the long run, it will only help Samsung to become a global leader in semiconductors, and it will be detrimental to TSMC.
Domestic semiconductor companies must increase their strength as soon as possible in order to seize the opportunity to cover up the vacancy
But this is an opportunity for domestic semiconductor companies. The current situation is that opportunities exist in front of and in the future, many customers need OEM technology and production capacity, and domestic semiconductor manufacturers need to catch up as quickly as possible, including SMIC, Hua Hong Semiconductor, etc., only to constantly improve their own strength, so that their industrial status has been Diverting forward can only make up for it when opportunities arise.
It may be possible to learn how TSMC and Samsung succeeded. In addition to the rigorous corporate culture, continuous refinement of technology, and deep-rooted customer service spirit, TSMC has left its competitors far in the past few years, winning 60% of the world's market share in one fell swoop. In fact, there is a secret.
TSMC's Process team has strong capabilities. In addition to the basic technology development capabilities and the DTP (Design Technology Platform) division, the original Open Innovation Platform OIP (Open Innovation Platform) is the secret weapon of Panding Peak.
Samsung's process capability and design capability are very strong, and its process capability is comparable with that of TSMC. In terms of design capabilities, Samsung will be able to make products such as its own processor Exynox. Samsung's design capabilities are therefore higher than that of TSMC, which is also Samsung's capability. The reason for the rapid emergence in the foundry field.
From the above-mentioned key point of view, the pace of domestic semiconductor giants, in addition to actively make advanced process technology such as 28 nm, 14 nm, 10 nm to make, whether to strengthen the process capability team and design ability team? Only by playing the basic skills can we go steadily and further. These seemingly details are important.
For example, SMIC does not have a complete team that resembles TSMC's Design Technology Platform (DTP). The reason for this is that such talents must use high-paying talents to scoop up their talents. The money will go down and talents will come. This is the same iron rule. Domestic Huawei is a good example.
Intel's semiconductor strength is still very strong, but it is obviously an elephant that cannot dance lightly, because even the progress of the new process is constantly being skipped. Although this does not affect the performance of each new generation of processors, it should not be the same as a leading company.
This time, Intel once again took the foundry business as the regulator of its main business, and its production capacity became scared and gradually faded away. If Intel wants to make a big splash in the foundry industry, when it calls on its customers, what kind of customer will it have? What's most important is that this kind of development will plant the seeds of South Korea’s semiconductor forces and influence the displacement of the global foundry industry. Intel may have his own wishful thinking, but the world can do its best. Some things are true and they can't go back.
The window of opportunity for the global foundry industry seems to be about to start again. This will be an opportunity for domestic semiconductor companies to grasp. Opportunities created by dramatic changes will be a key to rewriting history. .
The pin refers to the connection between the internal circuit of the integrated circuit (chip) and the peripheral circuit, and the pin constitutes the interface of the chip. According to the function, the pins of AT89S52 can be divided into four categories: main power supply, external crystal oscillator or oscillator, multi-function I/O port, and control, strobe and reset.
Terminal Pins,Terminal Hardware Pin,High Precision Terminal Pins,Terminal Pins For Pcb
Sichuan Xinlian electronic science and technology Company , https://www.sztmlchs.com