This article is from WeChat official account:Yuanchuan Research Institute (ID: caijingyanjiu), author: Chang Chen, Zhou Zhehao, support: far Sichuan Institute of Science and Technology group, the title figure from: vision China

After pressure from the United States, TSMC’s emergency contract manufacturing, and Huawei chartered flights, Kirin 9000 finally appeared alongside Huawei Mate 40. This is the strongest mobile phone chip in Huawei’s history, it is also Huawei’s first 5nm process mobile phone chip, and it is likely to become the last self-developed chip for Huawei mobile phones.

With Huawei’s consumer business in the global market, HiSilicon, which is responsible for the design and development of Huawei’s high-end chips, has also gained fame, and the Kirin series chips have once become the pearl of China’s semiconductor industry. But with the step by step pressure from the US government, the shortcomings in the key links of domestic chip manufacturing have been exposed.

This is not the first time Huawei has encountered political pressure. As early as ten years ago, Cisco CEO Chambers commented: “25 years ago I knew that our strongest opponent would come from China, and now it is Huawei[1].” Interestingly, the argument that “Huawei has financial support from the Chinese government” came from Chambers’s speech at a European summit.

In 2003, due to a Cisco lawsuit, Huawei’s core routing network equipment and other businesses were banned from entering the US market; in 2007, Huawei wanted to acquire 3com but was blocked by the United States; in 2010, Huawei extended specific assets to 3Leaf in California to acquire olive branches, the United States Stop here.

In the Obama era, Huawei encountered an investigation by the US government. The anti-China pioneer Marco Rubio we are familiar with is one of the important promoters.

Before Trump came to power, Huawei could still gain a foothold in Europe and Latin America. But the difference this time is that in the past it was only “not letting Huawei sell in”, but now it has “not letting Huawei makeIt’s really not too much to use the most difficult moment to summarize Huawei’s current situation.

HiSilicon, born out of the basic research department of Huawei, has grown from the marginal role of set-top box chips and video chips to a pivotal force in the semiconductor field. Recovering the rise of HiSilicon is also an aspect of observing Huawei’s glory and predicament.

From edge to center

In early 1993, in order to commemorate Huawei’s “survival”, Ren Zhengfei specially went to Hong Kong to customize 100 gold medals and rewarded 100 outstanding employees who persisted in the company’s difficult times.

Without a year, Huawei is facing a crisis again. At the window on the fifth floor of Shenzhen Nanyou Shenyi Industrial Building, 49-year-old Ren Zhengfei said to an engineer in a room: “The research and development of new products is not successful. You can change jobs. I can only jump from here! “

Ren Zhengfei’s new product refers to the C&C08 office switch. After Huawei made a breakthrough in customer switches, it turned its attention to office switches. However, things are not as simple as imagined. JK1000 failed miserably when it came out. By C&C08, Huawei’s situation is not optimistic.

The bank does not give loans. Ren Zhengfei can only rely on loan sharks. He even said: If anyone can get a loan of 10 million yuan, he does not have to go to work for a year and his salary is still paid.

Fortunately, C&C08 was a great success later, and it also opened the era of Huawei’s chips. In order to use its own developed chips for C&C08, Huawei established the Huawei Integrated Circuit Design Center in 1991. Ren Zhengfei hired Xu Wenwei to take charge of chip research from Yilida, a Hong Kong-funded enterprise not far from Huawei.

About this digging, there is a story that does not circulate very much—Elita was not very happy, and Xu Wenwei was in trouble.

In 1991, Huawei had the first ASIC chip with its own intellectual property rights. Two years later, Huawei’s first ASIC chip designed through its own EDA came out.

In 2004, Huawei separated its ASIC design center and established a wholly-owned subsidiary HiSilicon, also known as “Little HiSilicon”, responsible for the successful chip export business. The research and development of the system chip and the public platform are still under Huawei, responsible for the high-end chip attack, which is generally referred to as “Hai Si”.

In the early days of HiSilicon’s establishment, Ren Zhengfei set two goals, namely, recruiting 2,000 people and reaching 4 billion yuan in overseas sales within three years. The first goal was quickly achieved, but the second was far away. At first, HiSilicon made a SIM card chip. When the project was approved, there was ten dollars on the market. When it was made, it had fallen by one yuan. It felt like “Big Brother smuggled BB machines in jail for 15 years, and he was worthless when he came out.”

The turning point appeared at the TAIPEI COMPUTEX exhibition in 2006, where HiSilicon launched the H.264 video codec chip Hi3510.

At the time, the security giant Dahua was developing the second-generation DVR, but there were very few chips on the market that support H.264. As a result, the two parties started cooperation, and Dahua and HiSilicon had a contract with 200,000 H.264 video encoding chips, which gave HiSilicon to taste the sweetness.

Three years later, HiSilicon has made persistent efforts to launch SoC(system chip), which optimizes costs while surpassing the speed of research and development of peers, thereby attracting To the attention of Hikvision.

In the process of opening the chip market for set-top boxes, HiSilicon responded to an old saying that opportunities only favor those who are prepared.

HiSilicon’s set-top box chip was successfully taped out in 2007, but no one trusts this novice product, even his brother Huawei’s DVB(Radio and Television Digital TV) The set-top box also pays for it.

One year later, Guangxi Telecom and Huawei’s 100,000 set-top box order stalled because the chip solution provider “Broadcom” was unable to stock up in a short period of time. HiSilicon was ordered to meet the urgent need and stabilized in the set-top box chip market. Up the heel.

Huawei’s decision to take part in the mobile phone industry in person should have been a good opportunity for HiSilicon to make a big splash. Unexpectedly, Hisilicon dropped the chain when he should be performing.

In 2009, Huawei launched the first-generation mobile phone chip K3V1. It can be said that its manufacturing process was already behind when it was released. If the later K3V2 is the chip that made Huawei Hisilicon the worst scolded, then K3V1 is not even qualified to be scolded, and the Huawei C8300 with it is also uninterested.

Three years later, HiSilicon decided to take another gamble. It took the lead in Texas Instruments and Qualcomm and released theThe second-generation mobile phone chip with the smallest quad-core A9 architecture processor-K3V2. Huawei’s promotional video on K3V2 introduced it like this at the time: the world’s top quad-core processor, the perfect combination of high performance and low energy consumption. /p>

Unfortunately, things later developed in the opposite direction: K3V2’s disastrous design caused the phones that matched it-from Mate1, Ascend P2 to P6 and D2-to be complained and dissatisfied with users.

Nowadays, users who go to the archaeological video at station B will still use the barrage to complain: it is four fireballs.

From mobile phone to chip

2002 was a year when the domestic mobile phone market exploded. That year, ZTE mobile phone became one of its three strategic businesses from being on the verge of being sold, earning more than 10 billion yuan.

At the same time, foreign brands such as Ericsson, Nokia, and Panasonic have made huge profits by virtue of their cooperation with China Mobile. A Panasonic GD88 MMS mobile phone can be sold for more than 8,000 yuan, which is hard to find. [1].

Huawei is almost the only one absent. Ren Zhengfei has always felt that “making mobile phones” is the same as “doing real estate”. Even when employees proposed to establish a mobile phone project as soon as possible, Ren Zhengfei was outraged: “Huawei does not do mobile phones. , The conclusion has long been made, whoever talks nonsense, who is laid off!”

However, with the opening of the mobile phone market and the various side effects of employees from all walks of life, Ren Zhengfei, who is accustomed to thinking through “criticism and self-criticism”, finally agreed to hold a project discussion on mobile terminals.

After some discussion, Ren Zhengfei said to Ji Ping, who is in charge of financial affairs, “Ji Ping, use one billion to make mobile phones.” Then he said to the others in the room, “Being a mobile phone is not the same as making a system device. The approach and play are different. Huawei has to set up an independent terminal company to make mobile phones and operate independently! You guys plan how to do it[2] .”

In November 2003, Huawei Terminal Company was formally established. In the following six years, Huawei’s mobile phone business was mainly to make OEMs for operators, which can only barely be called the start-up stage. And do postThe card machine was hard and tired yet not making money, and the mobile phone was once reduced to a fringe business.

At this time, Huawei’s mobile phones are presented in two ways: either sell them or transform them completely.

Immediately afterwards, the outbreak of the iPhone completely changed Huawei’s thinking. The adaptation brought by Apple not only brought explosive growth to Huawei’s communications business, but also strengthened Ren Zhengfei’s determination to become a mobile phone.

In December 2010, Huawei Mobile held a high-level symposium. At the meeting, Ren Zhengfei upgraded the mobile phone business to one of the company’s three major business segments, shifting the focus of products from low-end OEMs to consumer-centric It is a high-end independent brand, and is proud to be the world’s first [4].

From 2B to 2C, the difficulty can be imagined. After pursuing the ultra-thin Ascend P1 and the heating machine Ascent D1 equipped with the self-developed chip k3v2, after being beaten by the market one after another, Huawei began to truly shift from an engineer’s straight male thinking to a consumer thinking: replacing industrial parameters with beauty .

Huawei Ascend P1

Yu Chengdong, who is in charge of the mobile phone business, devotes all resources to desperately, and he is pressing on one product: Huawei P6. In terms of product technology and quality, nearly 1,000 R&D personnel have been invested. For the metal battery cover alone, Huawei’s supplier has trial-produced 1 million pieces before finally daring to mass produce it[4].

In terms of brand building, Huawei hired Yang Zhe, the boss of Samsung’s China Brand Department, and upgraded its advertising slogan to “beauty is an attitude” and “flowing years.”Brand.

P6 is priced at 2688 yuan, breaking through the mid-range price, and finally sold 4 million units. The best performance of this Huawei mobile phone is 1 million units, and the P6 is a great success. After this battle, Huawei truly gained a foothold in the mobile phone market.

However, behind the success of P6, Huawei has a more widely criticized pain point in its mobile phone business-HiSilicon’s K3V2 chip.

Starting from the D series in 2012, the P series and Mate series were equipped with HiSilicon K3V2 chips. Until P6, even Yu Chengdong was shaken, and Ren Zhengfei still insisted on using it. When the old people from the Shanghai Institute mentioned P6, they couldn’t help but complain. The performance of the chip K3V2 was poor, and the problem of repeated recurrences in the final mass production of the mobile phone was not resolved until it was launched. The R&D personnel are miserable.

At this time, consumers’ aversion to K3V2 has reached its extreme. Netizens wrote a couplet to ridicule Yu Chengdong: “Hisilicon is forever, a ( K3V2)Forever spread”. The question is, why does Ren Zhengfei insist on using garbage K3V2?

Ren Zhengfei’s principle is “eat your own dog food first”-consumer chips must be continuously used in real scenes to expose problems, and only when problems are found can targeted medicines be corrected.

Without Huawei mobile phones, HiSilicon chips cannot grow up. This is an opportunity for Huawei compared with other mobile phone manufacturers, and it also contains huge market risks: Miss the window period for conquering cities and lose users.

The twenty-third article of Huawei’s Basic Law: We adhere to the “pressure principle”. We allocate resources with a strength greater than that of our main competitors on key success factors and selected strategic growth points. The earth concentrated its human, material and financial resources to achieve key breakthroughs.

In Ren Zhengfei’s mind, the HiSilicon chip is a long-term strategic investment for Huawei’s mobile phones. It must be focused on aggressively until it wins Shangganling.

From Kylin to Kylin+N

On May 17, 2019, HiSilicon President He Tingbo sent an open letter in response to Huawei’s inclusion in the U.S. Department of Commerce’s Bureau of Industry and Security (BIS) The list of entities.

He Tingbo said in the letter that Huawei had long expected that the United States would embarrass Huawei one day, and HiSilicon’s existence is to take precautions: “For this assumption that it will never happen, thousands of HiSilicon children have embarked on technology. The most tragic Long March in history is to create a’spare tire’ for the company’s survival.”

“These efforts have been integrated, turning the tide, ensuring the strategic safety of most of the company’s products, and the continuous supply of most of the products! Today, this darkest day is everyone’s The day when ordinary children become heroes of the times!”

According to what Tingbo said, HiSilicon has made arrangements in all aspects to prevent its own terminal products from being “stuck” by a certain aspect of chip technology in the future. Among them, HiSilicon has devoted the greatest effort to Kirin chips.

In 2013, Huawei P6 S was unveiled with the new Kirin 910 chip; in May 2014, P7 was released with Kirin 910T; until September 2014, Mate 7 equipped with Kirin 925 came out, which became the Feng Shui that determines the fate of Huawei’s mobile phone business ridge.

At that time, in the competition between Apple and Samsung, and no domestic mobile phone dared to bid up to 3000 in the market, Mate 7 decided to challenge the high-end market.

However, at the mobilization meeting of each representative office of Mate 7 in China, the product manager called out a sales target of 1.2 million units, but there was only one article in silence. Everyone was not optimistic. The colleagues of Huawei’s largest domestic representative office came to the stage to claim the sales target. At that time, he stretched out a finger and only shouted “10,000 units” [5].

As a result, Mate 7 has been on the market for only a week, and various regions have been sold out to request replenishment. Finally, global sales exceeded 7 million units in one fell swoop. Huawei has won its exclusive right to speak in the high-end machine market.

In addition, in April 2016, Huawei P9 equipped with Kirin 955 cooperated with Leica and began to promote the camera function. In order to achieve a smooth dual-lens experience, the Kirin 955 chip deliberately introduced a new image signal processing chip. , Huawei’s P series, which focuses on taking pictures, began to make an appearance.

It has been iterating to the Kirin 980 chip of Huawei P30, and it has been able to pk industry leader Qualcomm’s Snapdragon 855. From being scolded for thousands of years to becoming a world-class chip, HiSilicon finally fulfilled Ren Zhengfei’s expectations: We must stand up and reduce our dependence on the United States.

In August 2018, Huawei released the Kirin 980 chip

The success of the Kirin chip is inseparable from the cooperation of the baseband chip. In this business, the Balong (Balong)The chip has been deeply cultivated for many years.

From 2007 to 2010, HiSilicon cooperated closely with Huawei, and started the intensive 4G LTE R&D layout when the market was still in the 3G era.

Finally, in early 2010, Huawei launched the industry’s first terminal chip “Baron 700” developed by HiSilicon to support TD-LTE, supporting LTE FDD and TD-LTE dual mode. On the day of the first successful tape-out, the Shanghai Research Institute was full of joy.

More importantly, Barong broke Qualcomm’s monopoly on the baseband chip market and further improved its self-research capabilities through cooperation with Kirin chips.

Since then, HiSilicon has made persistent efforts. In 2012, it released the baseband chip Balong 710 integrated in Kirin 910, achieving a downlink rate of 150 Mbps with a year ahead of Qualcomm and MediaTek; a year later, Balong 720 came out with a downlink rate. Doubled; in 2015, Barong 750 was released, and the rate doubled again, and surpassed the MDM9x45, a product of the same period of Qualcomm, which is the “top market”.

After 2018, the frequency of HiSilicon’s product releases has increased significantly.

First, Huawei released the self-developed HiSilicon “Lingxiao” series of chips, positioning “home high-speed intelligent interconnection” for use in a variety of router products. One year later, the “Kunpeng 920” server chip integrated the four major structures of network, storage, main control chip and CPU in a single package.

Following up, the world’s first 5G base station core chip-“Tiangang” is alsoFreshly baked. Using this chip, the size of the base station is reduced by more than 50%, the weight is reduced by 23%, and the installation time is only half of the standard 4G base station.

The “Shengteng 910” unveiled in September 2020 is known as the “most powerful” AI processor. Huawei’s rotating chairman Xu Zhijun said that the target of the Shengteng 910 is the AI ​​computing power chip of Google and Nvidia. From what he said in an interview after the press conference, people can also feel Huawei’s high hopes for HiSilicon.

In the interview after the press conference, he attributed the reason for HiSilicon’s success to Huawei’s daring investment: “Last year’s financial report was released. Some voices said that Huawei did not earn much from Tencent and Ali, but Mr. Ren (Ren Zhengfei) criticizes that we are still earning a lot, indicating that the strategic investment is not enough, and we need to write various reviews.”

Kirin, Shengteng, Kunpeng, Barong, Tiangang, Lingxiao, Honghu… HiSilicon’s “Guofeng Chip Family” is expanding at an average update rate of half a year, and it is compatible with Huawei’s entire product line. Complement each other.

For more than ten years, HiSilicon has always adhered to the “old vow”: to stick to Huawei, where is the lack of “core” to make up.

Ending

Ren Zhengfei once stated exactly the importance of Huawei’s chip research and development: “During the siege, the team was arranged vertically. The siege troops were concentrated to tear open a hole, and then the two main forces entered through the hole. , Expand to both sides. Then enter four more divisions, deepen and expand the results to both sides.”

For Huawei, the chip business is an important and sharp force to attack the market and concentrate on breaking through the market. When chip technology makes a breakthrough, other product lines can also achieve large-scale market victory.

At the 2020 Summit of the China Information Technology Association in August, Yu Chengdong said regretfully: “Huawei has been developing for more than ten years. It has been a difficult process from being seriously behind to a little behind and then catching up to lead We have suffered a great loss.”

However, Huawei has not given up on chip research and development. Yu Chengdong called on the domestic semiconductor industry chain to strengthen cooperation and explore ways to manufacture semiconductor products under pressure from the United States.

Ren Zhengfei visited four colleges and universities intensively before and talked about the importance of basic scientific research several times. It seems that it also means to accumulate talents for Huawei’s next step.

A little review of business history will find that many companies that have successfully comeback did not fight hard on the leader’s territory, but first quietly accumulate technology, save the fire, survive the cold winter, and adopt new technical routes. Achieve breakthroughs and overthrow those with vested interests.

Huawei’s core manufacturing was not a whim, let alone a whimsical one, but a prudent plan for the future. Huawei is now immersed in the layout of the industrial chain, while relying on Hongmeng to do ecology in the AloT field, naturally it is not pursuing immediate quick victory. Instead, stay at the poker table, accumulate strength, and wait for breakthrough opportunities.

For Huawei, except for victory, there is no other choice.

Reference material:

[1] The original Huawei employee proposed to make a Huawei mobile phone 15 years ago. Ren Zhengfei was furious and looked up.

[2] The past of Huawei’s mobile phone: the story of the rise of a hardcore straight man, the boss of the meal.

[3] Made in Shenzhen: Huawei and Foxconn in the historical process, the boss of Fantongdai.

[4] Huawei terminal strategy, Rui Bin, Xiong Yuejia.

[5] Six years and seven stories of Huawei terminals, daily science and technology.

This article is from WeChat official account:Yuanchuan Research Institute (ID: caijingyanjiu) author: Chang Chen, Zhou Zhehao, support: far Sichuan Institute of Science and Technology group