Reborn in 2011, 985 missing from two copies

Chapter 456 Global layout, sugar-coated bullets for TSMC
To put it bluntly, all results speak for themselves.

When doing scientific research in universities, the candidates are either promoted or left.

No matter whether you have a PhD from a 985 university or a C9 university, if you fail to meet the assessment requirements and produce results, you will be fired.

The same is true in business.

Even if it is Pang Liguo, if the development of the Wing Loong 3000 baseband fails, it will also lead to the failure of the Kunpeng 700 and Kunpeng 702 SOC, and Wang Yi's billion investment will be wasted.

Then Wang Yi will also let him go!
In the semiconductor industry, there is no second chance!
Nie Zheng has been in the business for many years and is very familiar with these unspoken rules. He immediately expressed his opinion:
"Mr. Wang, don't worry. I will go all out and make a 28-nanometer CPU by the end of next year!"

"Okay, I'll wait to celebrate for you!"

Wang Yi spoke calmly, and next, it would depend on the performance of each R&D center.

We now have several R&D centers, and all three major businesses, CPU, GPU, and AI chips, are invested in.

However, subsequent financial support will depend on the performance of each R&D center.

To put it bluntly, it is about raising insects and internal competition.

Competition between different R&D centers in the same business unit.

There is also competition between different business units.

R&D centers with outstanding results will increase investment, upgrade their strategic level, and receive generous year-end bonuses.

R&D centers with mediocre results will have their funding support reduced and will be put on the waiting list for assessment, so they can forget about year-end bonuses.

As for R&D centers with poor or even no results, it is possible to change leaders, cut salaries and demote positions, lay off employees, or even cut the entire business unit.

Although VIA was previously a semiconductor retirement home, Wang Yi would not let them really retire.

Wang Yi acquired VIA and invested heavily in research and development in order to achieve results and make money.

As long as you can produce results, treatment and status will not be a problem.

But if they spend money and don’t get any results, Wang Yi will let the other party know what capitalist tactics are!
Afterwards, Wang Yi approached Williams again and said, "William, the CPU R&D Center and NPU R&D Center in the imperial capital are both in your hands. You are also responsible for the overseas Centaur team."

After all, NPU is integrated into the CPU SOC just like the integrated graphics card.

The current computer's CPU SOC includes CPU and integrated core graphics, and will also integrate NPU in the future.

For example, in the GPU department, if a powerful GPU is developed, Wang Yi will launch it separately.

If the performance of the GPU developed is too poor, Wang Yi would be embarrassed to launch it as a discrete graphics card, which would be too embarrassing.

At that time, it will be directly packaged together with the CPU to make the CPU's core graphics card.

Just like Intel's third-generation i5 next year, it will have integrated Intel HD Graphics 2500.

What level?

It is roughly equivalent to the GT210 on a desktop, an entry-level independent graphics card.

It is no problem to use it for office and study, but as for games, you can only play 2D games, such as CF, DNF, etc., don't even think about 3D games.

Similarly, if S3 graphic is powerful and the CPU it produces can catch up with Nvidia's mid-range graphics cards of the same generation, Wang Yi will mass-produce it and make it into an independent graphics card.

But if it is only equivalent to an entry-level low-end graphics card, it can only be made into the CPU's core graphics card.

If they can't even make an entry-level graphics card, then there's no need for S3 graphic to exist.

"Okay, Chairman, I'll take care of everything." Williams said solemnly.

One CPU R&D center headquarters, two NPU R&D centers, three core departments were handed over to him at once, which shows Wang Yi's trust in him.

There is no way, Wang Yi doesn’t know much about the overseas team.

Even if other people take charge, they won’t be able to control the situation.

Only Williams can play.

Williams developed the Kunpeng 702 SOC in such a short period of time, and achieved the feat of 40 nanometers beating Qualcomm's 28 nanometers. He became a legend in one fell swoop and has become a top player in the semiconductor field.

With him in charge, the overseas Centaur teams have nothing to say.

The performance is there.

Who dares to be arrogant unless they make a CPU that can defeat Intel?

But unfortunately, the current Centaur team no longer has such ability and can never catch up even in their dreams.

Then you can only obey Williams.

Another consideration is that Wang Yi trusts Williams.

Williams is not a power-hungry person; instead, he appoints people based on their merits.

People like George and Zhang Chao were recommended and promoted by Williams, and now they are both capable of leading their own businesses.

If it were someone else, they would never let George and Zhang Chao, their capable subordinates who might pose a threat to them, to come forward. Any chance of coming forward would be completely stifled!

Moreover, Williams has strong ability to withstand pressure, is good at leading the team to open up new areas, and is a qualified marshal.

Wang Yi will naturally make use of him!

"As for the GPU R&D center in Seoul (Wuhan), it will be temporarily handled by the current technical director. As the general manager of Xingyi Semiconductor, you should do a good job of the assessment. If that person is unqualified, we will replace him!" Wang Yi instructed.

"Yes!" Williams responded, only to feel the burden on his shoulders become heavier again.

Now Xingyi Semiconductor has a mobile phone SOC division, a mobile phone baseband division, a comprehensive chip division, a computer CPU division, a GPU division, and an AI chip division.

Each business unit has its own leader and its own R&D director.

But the head of the entire Xingyi Semiconductor is still Williams!
Wang Yi doesn't have time to keep an eye on it every day, so it is Williams who keeps abreast of the progress of each business unit and R&D center.

Grasp the assessment and talent discovery.

For example, Williams would recommend the core talents of the Centaur team, the VIA team, and the Seoul GPU team to Wang Yi once he discovered outstanding employees.

Each department should recommend at least three to five people.

Wang Yi then arranged for people to go through layers of assessments and selected the most suitable person to become the new leader.

In the following days, VIA's core semiconductor talents continued to fly back to the capital from Taiwan.

There are also core Centaurs talents flying to the imperial capital from overseas.

Of course, the latter is less.

After all, the Centaur team is not like the previous mobile chip team of Texas Instruments.

Texas Instruments is abandoning its mobile chip business and cutting its entire division.

George and others were about to lose their jobs, and couldn't find better treatment, so they had no choice but to come to the imperial capital and join Xingyi Technology.

To put it bluntly, George and his team came to the imperial capital and joined Xingyi Technology. It was somewhat forced and they had no choice.

But the Centaurs are different. They are very strong and can find a good job even if they leave and go to Intel or AMD.

Therefore, Wang Yi could only adopt a conciliatory policy.

Use high salaries to lure those who are willing to come to the capital. Give them high salaries.

Those who do not want to come can continue to work in the overseas Centaur team, where they will receive adequate remuneration and do not have to resign.

After all, the Centaur team is strong and capable of developing AI chips, and they have a bright future.

Wang Yidu treated it as a treasure.

Even if we don't come to the imperial capital now, we have to keep them stable.

Subsequently, they will be screened and evaluated based on their abilities and performance. Young people with outstanding performance will be transferred to the headquarters with the excuse of promotion and salary increase!
Well, if you stay in the overseas branch, promotions and salary increases will be slow.

Come to headquarters and get an immediate promotion and salary increase.

Wang Yi’s main point is that his virtue can convince people and his humbleness can make people approachable!
Jian Ziyan had already arranged reception and apartments for these incoming semiconductor talents.

After finishing the work, Jian Ziyan approached Wang Yi and said, "Chairman, we have arranged for the talents coming this time, but there will be several more batches of talents coming in the future. I have a suggestion. It is not appropriate for us to always rent apartments. How about buying a few apartment buildings in the name of the company and using them as dormitories for our talents?"

"Buy an apartment building in the name of the company?" Wang Yi's eyes lit up:
"That's a good idea. Xingyi Semiconductor keeps poaching people, and the big names we recruit need to find accommodation. We might as well just buy a house. But we won't buy an apartment because the tenants are too mixed. We'll buy a small house!"

Housing prices in the capital are high, and most houses are small in size.

There is no way, housing prices are too expensive and few people can afford large apartments.

When buying a house in second- and third-tier cities, you can either buy a small three-bedroom apartment of 100 square meters or a large three-bedroom apartment of 120 square meters.

But when buying a house in the capital, you usually buy a two-bedroom apartment of 40 to 50 square meters, or a small three-bedroom apartment of 70 to 80 square meters.

As for a house over 100 square meters? Too expensive, too extravagant.

At present, VIA employees coming to the capital need to solve the housing problem, and the big names and talents recruited later also need to solve the housing problem. Before, they all rented apartments and houses.

Might as well buy it now.

A few years later, housing prices in the capital skyrocketed. Even in Zhongguancun, the annual rent for a 50-square-meter house would be more than yuan.

Currently, a 100-square-meter house is only priced at 150 to million yuan.

The rent in the next ten years will almost catch up with the current house prices. There is no reason to just rent and not buy, and then work for the landlord.

Many years later, whether it was demolished or not, people made a lot of money.

It is easy to buy a house in the name of a company.

"Zi Yan, just go ahead and buy some houses. Buy the houses of those landlords who have many houses and use them as apartments for your employees. If you are short of money later, you can just mortgage them to the bank for a loan," Wang Yi instructed.

They only target landlords who own many houses. These people are too lucky and have not had to struggle for several generations.

The rent collected every month is higher than the income of elites in big companies, and there is no need to pay taxes!
No matter how many houses Wang Yi buys, it will not affect their quality of life. After all, as long as they don't do anything stupid, they will have endless money to spend.

It doesn't matter for ordinary people, but they should also be given the possibility of becoming multi-millionaires through demolition.

"Okay, Chairman, I'll do it right away."

After Jian Ziyan left, Wang Yi began to check the financial statements of each business unit.

With the addition of VIA talents, the CPU business unit of Didu has developed rapidly and has a large number of employees and generals.

Coupled with VIA's chip technology and foundation, it is only a matter of time before they develop a CPU.

The VIA headquarters in Taiwan also burst into new vitality due to Wang Yi's capital injection.

Although there were fewer people, we also cut off the original messy businesses and started to focus on CPU research and development.

Even the S3 graphic R&D team in Seoul has started recruiting people and focusing on developing a new GPU after receiving financial support!

There is no way. GPU is too difficult and it is unrealistic not to recruit people.

In the simplest case, a CPU may be produced in one year.

But GPU is estimated to take two years, or even longer!

Fortunately, we have S3 Graphics' background and technology, otherwise it would have been difficult for two years. After all, S3 Graphics also made several 3D graphics cards in the past few years, but they were much worse than NVIDIA.

But whether the final product can be made into an independent graphics card or can only be used as the core graphics card of the CPU depends on their capabilities.

But Wang Yi had said the ugly truth in advance.

Whether the S3 graphic division can become an independent GPU team or be a subsidiary of the CPU department to produce integrated graphics depends on their results.

If a powerful independent graphics card is produced, the S3 Graphics Division will officially become the graphics card department of Xingyi Semiconductor, with the same SSS-level strategic positioning as the CPU department, and all treatment and status will rise.

But if it can only produce entry-level graphics cards that can only be used as core graphics cards, then the S3 graphic business unit will also become a group of the CPU business unit.

The treatment of the formal GPU department and a small group in the CPU department is a world of difference.

Once the group becomes part of the CPU department, the entire S3 graphic team will be broken up, with some people joining the CPU integrated graphics group.

Others who are capable will stay and join the new graphics card department.

Those who are incompetent will be fired directly.

The N+1 compensation is in place, but the employee has only been employed for one or two years, so at most it’s three months’ compensation, not much.

There is no way. Semiconductors are too expensive, and Xingyi Semiconductor will not keep waste.

Under this situation, the entire S3 graphic team began to sprint and overcome difficulties with all their strength.

The top leader, Yu Longxing, the former R&D director of S3 graphic, is very hands-on and wants to live and eat in the company.

The top boss works day and night like this, and other employees dare not be lazy either.

In addition, the salary Wang Yi offered was 30% higher than before. If he could achieve more results, the year-end bonus would be even more generous.

For a time, the entire graphics card department was full of fighting spirit and united as one.

Similarly, the overseas Centaur team was also in high spirits and ready to work hard after learning that Xingyi Technology had become their new owner and that their salaries and R&D funds were guaranteed.

When they learned that they were going to develop AI chips, everyone was a little confused.

But through several video conferences, Wang Yi and Williams personally explained the development direction of NPU, and they all became interested.

I feel like this is a challenge and a new track!
Let the Centaurs create another glorious track!
Wang Yi welcomes the Centaurs' working attitude.

After all, AI chips are just in their emerging stages, and everything is done by trial and error.

The Centaur team is located overseas and is more suitable for recruiting global elites to join.

Of course, the entire research and development process is carried out together with the domestic NPU headquarters.

All the results and progress are also under control in China.

In fact, most IC design companies have multiple R&D centers around the world.

This is true whether it is Qualcomm, MediaTek, or HiSilicon.

Today, Xingyi Semiconductor has also entered a global layout and multiple R&D center model.

Williams and Nie Zheng were in charge of VIA's affairs. The next day, Wang Yi came to the mobile phone department and asked, "Mr. Zhu, how is the mobile phone business going recently?"

The mobile phone is the basic infrastructure and no problems can occur.

"Chairman, mobile phone sales are booming, especially the Kunpeng version, which is selling well across the board."

Zhu Changlin was a little excited: "The Kunpeng version of xphone 2 was launched on the 24th, and the spot supply was supported until the 27th, then pre-sale started and delivery started within four days. In fact, our Kunpeng 702 SOC production capacity crisis was completely resolved in those days, and the Kunpeng version of xphone 2 was quickly put into production. It was fully available at the end of the month."

“Only out of stock for three days? That’s good.” Wang Yi breathed a sigh of relief.

Kunpeng 702 was out of stock for only three days, which is considered very good.

像是iPhone 15，9月22开售当天下单，官网买10月22发货。

When iPhone 5s was released, many people placed an order on the first day and waited two or three months for it.

This is called out of stock.

Xingyi Technology was out of stock for only three days, and the product was shipped within three days, so it really wasn’t out of stock.

"How about the Xingyi Tablet X1 Ultra?"

"The X1 Ultra tablet was similar, it was out of stock for about three days, but it was back in stock in China by the end of the month. In December, it was back in stock worldwide. The Kunpeng version of the Xphone 1 Pro was also out of stock for four days, but was back in stock at the end of last month. The impact was not significant."

"That's good." Wang Yi breathed a sigh of relief and began to check the sales carefully.

The one with the highest improvement is still the Kunpeng version of xphone 1pro.

The average daily sales volume in the past week has remained stable at around 40 units. Even if the volume drops a little later, the sales volume in December will still exceed 12 million units.

Logically speaking, Xphone 2 has been released for three months, and the monthly sales of Xphone 1 Pro should drop below 500 million.

But because of the chip upgrade, xphone 1pro really has a second spring.

There is no way. After all, the Kunpeng 700 SOC is a huge upgrade compared to Nvidia Tegra 3, and it is not an exaggeration to say that it is a whole new generation.

In addition, the xphone 1 pro is one thousand yuan cheaper than the xphone 2, so many people are willing to buy it.

The sales volume of Kunpeng version of xphone 2 is still terrifying.

The average daily sales volume has stabilized at around 50 units, and the total sales volume in December is expected to exceed 12 million units.

As for the Xingyi Tablet X1 Ultra, it beats the fourth generation of iPad. It went on sale globally with daily sales of 30 units, and sales in December are expected to exceed 12 million units.

就连搭载鲲鹏700的星逸平板X1 pro，单日销量都有20万台，12月卖个400万不成问题。

Wang Yi calculated that Xphone 2 + Xingyi Tablet X1 Ultra could sell 12 million units in December, which would require 2000 million Kunpeng 2000 SOCs.

The xphone 1pro + Xingyi tablet X1 pro can sell 12 million units in December.

The current monthly production capacity of Kunpeng 702 SOC and Kunpeng 700 SOC is about 1500 million units, which is obviously not enough.

After all, in late December, the xphone 12 and Xingyi tablet X506 equipped with Kunpeng 1 will be launched, which will require tens of millions of Kunpeng 1 SOCs in total to castrate the algorithm.

Not to mention the release of borderless mobile phones at the end of January and the beginning of February, which requires at least tens of millions of Kunpeng 500s or tens of millions of Kunpeng 700s to be processed for algorithm casting every month.

Taking into account the decreasing sales volume of subsequent models, approximately 2500 million Kunpeng 700 SOCs will be needed every month!

Starting from February, 200 million Kunpeng 702 SoCs will have to be delivered to HTC every month.

The total production capacity of the two wafer fabs is only 3000 million pieces, which is not enough.

In response, Wang Yi looked at Zhu Changlin and said, "Old Zhu, the chip production capacity is not enough. We have to find a foundry."

"Then contact TSMC to manufacture Kunpeng 700?" Zhu Changlin immediately understood what Wang Yi meant.

"Yes! This month, Xingyi B1 wafer fab is still fully committed to the production of Kunpeng 702 SOC, with a mass production of more than 1500 million units."

“M1晶圆厂的部分生产线也开始生产鲲鹏702 SOC，月产500万。剩下的1000万产能继续生产鲲鹏700 SOC。”

"In addition, we will let TSMC manufacture the Kunpeng 700 SOC, with a monthly production of 1000 million units. If it is not enough, we will place additional orders later!"

In this way, 4000 million Kunpeng 702 SOCs and 4000 million Kunpeng 700 SOCs can be mass-produced in the next two months, which should be enough.

After all, as time goes by, the sales of Xphone 2, X1 Ultra, and other mobile phones are decreasing month by month.

By January, perhaps all 2000 million Kunpeng 702 SOCs may not be used, and the remaining inventory will be just enough to be delivered to HTC in February.

There is also Kunpeng 700, which will be dynamically adjusted based on market demand and sales of each product line.

It's all easy.

By March and April, the sales volume of Kunpeng 700 SOC may fall below 1500 million.

The extra Kunpeng 700 SOCs produced at that time can be sold to other friendly companies.

There is no way. The 28nm process is not that easy to master, and the Kunpeng 902 SOC still has to be manufactured by TSMC.

By then, Xingyi Semiconductor's two major wafer fabs will inevitably have idling 40-nanometer production capacity.

Instead of wasting the production line, it is better to produce more Kunpeng 700 SOCs and sell them to make money.

With Kunpeng 902 SOC ready to go, even Kunpeng 702 SOC can be sold in large quantities in March and April before Snapdragon 800 goes into mass production.

"Okay, Chairman, I will contact TSMC right away to manufacture the Kunpeng 700 SOC."

Zhu Changlin responded and went to arrange it immediately.

Ten minutes later, Zhu Changlin returned again: "Chairman, it's done. TSMC gave us a good price. The total price of Kunpeng 700 SOC is 27 yuan per piece."

"Okay." Wang Yi was quite satisfied.

The Kunpeng 700 SOC is simpler and less expensive than the Kunpeng 702 SOC.

It costs 23 yuan for Xingyi Semiconductor to produce it itself, while TSMC costs 27 yuan, which is not expensive.

The main reason is that there are fewer orders for 40nm, and many of TSMC's 40nm production lines are idle.

"Changlin, please arrange the various factories and product lines."

"Don't worry, Chairman, nothing will go wrong."

Wang Yi nodded and turned to leave.

TSMC’s headquarters, however, has a completely different atmosphere.

"Chairman, Xingyi Technology has built a wafer fab and achieved mass production and packaging and testing of 40-nanometer technology. It is considered an enemy of TSMC. Why should we still accept their orders? Help them produce chips?"

Lao Zhang snorted coldly: "Stupid!"

"So what if Xingyi Semiconductor has a 40-nanometer process? If they don't have enough production capacity, won't they still have to come to us? The previous smart home chips and router chips were ordered in quantities of million or million each time."

"Last month, the Pterosaur 3000 baseband started with 500 million units, and quickly increased to 2000 million units!"

"Now there are another 1000 million Kunpeng 700 SOCs, and in the next few months, there will be 1000 million per month."

"We have so many orders coming to our door, but we don't want them. Samsung Semiconductor and Intel Semiconductor are watching us eagerly!"

"You want to give away all these orders?"

"I..." The subordinate was speechless for a moment.

Old Zhang shook his head: "Xingyi Semiconductor is a competitor, and also a major customer of TSMC. The previous cooperation has gone well, so all the extra orders of Xingyi Semiconductor are given to TSMC. It was so in the past, it is so now, and it will be so in the future."

"Don't forget, Xingyi Semiconductor only has 40-nanometer technology and cannot make 28-nanometer chips. Their 28-nanometer chips will have to be outsourced to TSMC."

The subordinate suddenly realized: "So that's it, I was too abrupt."

Lao Zhang said meaningfully: "If we have a falling out with Xingyi Technology, everything will be over, and the loss will outweigh the gain. It's better to just treat them as customers and make as much money as possible."

"Wang Yi didn't give these orders to Samsung or Intel, but only to TSMC. That's what he meant! Cooperation is a win-win for both sides!"

"Not only should we not break off relations with Xingyi Technology, nor blockade and suppress them, we should also serve them well and maintain a good relationship. A bigger picture will lead to a win-win situation!"

"Chairman, great idea!" The subordinate praised, but he was muttering in his heart:
'Why do I feel like this is Wang Yi's sugar-coated bullet?'

(End of this chapter)