Reborn in 2011, 985 missing from two copies

Chapter 573: Backstabbed, TSMC desperate?
These days, Huawei's self-developed Kirin chips are still in their infancy, accounting for only 30%-40% of Huawei's total chips.

Wang Yi changed the subject: "Huawei used Qualcomm chips and MediaTek chips before. You can also talk to them about selling our Kunpeng chips to them and try to reach in-depth cooperation. If necessary, you can give them some discounts in chip foundry."

"Okay, this is a win-win situation!" Zhang Chao agreed.

After all, Huawei buys chips from MediaTek and Qualcomm, and also buys chips from Xingyi Technology.

"Yes, it's a win-win situation. We win twice! Once by manufacturing Kirin chips and once by selling Kunpeng chips to Huawei." Wang Yi agreed.

Zhang Chao's mouth twitched: "..."

Win-win…win twice…great logic!

But if you think about it carefully, there's nothing wrong with it.

Afterwards, Zhang Chao acted decisively and set off that day, flying directly to Shenzhen Stock Exchange.

As for the research and development of Kunpeng 906 SOC, all verification work has been completed.

Currently, the wafer fab's R&D plant is preparing for tape-out, just waiting for tape-out at the end of the month!

These preparatory works are mainly related to chip production. It won't be a big problem if Zhang Chao is not there. He can just be there when the chip is taped out.

At the same time, Huawei also held a high-level meeting to discuss related matters.

Lao Yu was the first to speak: "Regarding the chip foundry issue of Kirin 910, there are currently two options. One is TSMC, which was originally planned. Our K3V2 is manufactured by them."

"But recently Xingyi Semiconductor has extended an olive branch to us, wanting to do OEM for us. And the price is more favorable."

“So how do we choose?”

Everyone fell into deep thought, with different thoughts in their minds.

"If you have any ideas, just tell us." Yu Dong said with a smile.

Everyone responded and spoke:
"I suggest using TSMC as the foundry. As a semiconductor giant, TSMC has more experience and more mature technology."

"I object. Although Xingyi Semiconductor is an emerging semiconductor company, it has made rapid progress and has a bright future. Their 28-nanometer process is excellent and is no worse than TSMC."

"Xingyi Semiconductor's 28nm process is very good, but the question is what's next?"

"TSMC and Samsung are both developing 20-nanometer processes, and will be able to advance 20-nanometer processes next year. We use TSMC for foundry, and we can also use 20-nanometer processes in the future, but what about Xingyi Technology? They just achieved 28-nanometer processes this year, and 20-nanometer processes are still a long way off, right?"

"This is a big problem. Once we abandon TSMC and choose Xingyi Semiconductor, it will be equivalent to completely offending TSMC. It will be a foolish dream to want to use TSMC's advanced technology in the future."

“We have to think deeply about this.”

"You think too much. Even if TSMC uses 20nm technology next year, it will not be Huawei's turn. Even Qualcomm cannot use it. TSMC's 20nm technology next year will be very limited, and all will be used for Apple A8."

"It won't be our turn next year. But the year after, there is hope."

"The year after next? How do you know that Xingyi Technology won't be able to achieve 20nm technology the year after next?"

"That makes sense. With Xingyi Technology's research and development speed, it is entirely possible to achieve 20-nanometer technology next year, so don't worry about it."

"The main reason is that TSMC's position is not firm. This time, for Qualcomm's sake, it directly broke the contract and stopped supplying Xingyi Technology with 40-nanometer Kunpeng chip orders. It was a super large order of 2000 million chips per month."

"We continue to manufacture for TSMC, but we cannot guarantee that they will not also cut off our chip supply, especially if our mobile phone sales surge in the future and surpass Apple, then anything is possible."

"That's right. What happened to Xingyi Semiconductor today is the future of HiSilicon Semiconductor. We have to think deeply about this."

"TSMC's position has always been unsteady. They use a lot of Xiaomi's technology and have to obey them. Xingyi Technology is completely different. They develop their own technology and have the final say on everything."

"But Xingyi Mobile and Huawei are competitors. Are you confident in handing over the orders to your competitors? Aren't you afraid of being strangled?"

Yu Dong waved his hand and said, "Don't worry about that. President Wang promised that no matter what happens, Xingyi Semiconductor will continue to manufacture chips for Huawei in a stable manner. TSMC dare not promise this and cannot guarantee it."

"And given Mr. Wang's character, he won't cut off the supply of chips to us. I personally think that Xingyi Technology can be fully trusted."

"Well said, at critical moments, it is our brother companies that we can rely on!"

A hearty voice sounded, and Mr. Ren walked in from outside the door:

"Although we compete with Xingyi Technology in the mobile phone business, it can be healthy competition, and we can make progress and develop together."

"In fact, Huawei phones, other domestic phones, and even global phones are now learning from Xingyi Technology and imitating Xingyi phones. In this respect, Xingyi phones are the teacher of all our companies."

"As for the supply cut, don't worry. If Xingyi Technology cuts off the supply, it's only because they don't want domestic sales anymore."

Yu Dong nodded repeatedly: "That's right. If Xingyi Technology stops supplying Huawei with chips, the Chinese people will be able to drown them with a single spit. They won't be able to sell Xingyi phones anymore."

Ren Lao smiled and said, "And Mr. Wang can't do such a thing, so don't worry about that. But it is very likely that TSMC will stop supplying. This time, the suspension of supply of Kunpeng chips to Xingyi Technology is a lesson for us. It is better to choose Xingyi Semiconductor."

"Seconded!"

"Seconded!"

Everyone nodded, and no one had any other opinions.

Just as Yu Dong and Ren Lao said, Xingyi Semiconductor will not cut off supply like TSMC, and this is the greatest strategic significance.

As for maliciously cutting off supply in order to suppress Huawei phones... unless Xingyi phones don't want to be sold in China anymore.

This kind of thing that triggers public anger will not happen.

The next day, Zhang Chao came to Huawei headquarters again.

Lao Yu warmly welcomed him: "Mr. Zhang, welcome, welcome. I knew you were coming, so I put aside all my plans, prepared tea, and waited for you."

Zhang Chao felt relieved, it seemed that the OEM business was settled, the key was how to negotiate the details!
"Thank you for your warm hospitality, Mr. Yu. I came here with sincerity this time."

"Sincerity? I'm really looking forward to it, hahaha." Yu Dong said with a smile, "To be honest, we have discussed the matter of Xingyi Technology's OEM internally. Some people support it, some people oppose it, but overall, there are more supporters. As for the OEM price, I don't know..."

Zhang Chao did not answer directly, but asked: "How many Kirin 910 will be mass-produced?"

"2000 million!" Yu Dong said decisively.

It was just as the chairman expected!
Zhang Chao continued: "There will be other chips next year, right? 2000 million is not enough for Huawei."

These days, Huawei's annual mobile phone sales are around 1.4 million units, with smartphones accounting for more than half, or about to million units.

Of course, feature phones also need chips, such as Qualcomm's dual-core, Snapdragon 200 series, MediaTek's single-core, dual-core, entry-level quad-core, Nvidia's Tegra 2, etc. These outdated processors have become processors dedicated to feature phones.

However, Xingyi Semiconductor is still blank in this entry-level processor.

哪怕是最差的40纳米3G芯片鲲鹏500，实际体验都和骁龙400不相上下，再加上64位的系统的加成，都快赶上骁龙600了。

There really are no dedicated chips for feature phones.

Xingyi Wafer Factory has limited production capacity, and Wang Yi looks down on entry-level processors, or feature phone processors.

That’s right, entry-level processors like Snapdragon 200 are designed specifically for feature phones.

The first generation of Redmi in the past was not even good enough, and at least it had to be a low-end processor at the Snapdragon 400 level.

Yu Dong did not hide it: "Yes, next year we will also release Kirin 920, Kirin 920, and Kirin 925, a total of three chips."

“What is the expected production capacity of these three models?”

Yu Dong understood that Zhang Chao wanted to give Huawei a quote based on Huawei's order volume!

This is all natural.

In the chip foundry industry, the typical scale effect has always been to serve different people differently.

It is natural that the larger the scale, the lower the OEM price.

Yu Dong said decisively: "Kirin 920 is the main product and will be used in Honor 7. The production volume should be around 920 million. Kirin 7 will be used in Huawei Mate 928, and the production capacity is also 500 million. As for Kirin 2500, which will be used in Honor Extreme Edition, the production capacity is expected to be 910 million. The total is about 4500 million! Together with Kirin , the total is million!"

Lao Yu is full of confidence, as if to say that our size is just this big and there is no water in it.

Zhang Chao didn't believe it at all, muttering in his heart: "Next year, Huawei's total smartphone sales will only be 50 to million, and the proportion of Kirin chip models is small, it is impossible to exceed %. Giving them a % discount will be enough."

'It's said to be 4500 million, but if you remove 3000% of the water, it's only about million.'

"The Kirin 910 is used in Huawei's P6S and Mate 2 flagship phones, as well as Honor's thousand-yuan phones, totaling about million yuan."

"As for Kirin 920, at most 7-925 million units. However, the Kirin of Mate may be a little more, million units."

"As for the Supreme Edition of Kirin 928, the Supreme Edition, the sales volume is even less, five million? At most one million!"

"So the total is 30 million, no more."

Zhang Chao figured out everything in an instant, but he didn't expose it:

"President Wang has explained to Huawei that he has given special support. Logically, the foundry price of Huawei's Kirin 910 is calculated based on the 910 million yuan large volume of Kirin 920. The subsequent foundry price of Kirin 920 is calculated based on the small volume of Kirin ."

Yu Dong nodded, these are the market conditions in the industry.

For example, a 28-nanometer high-end mobile phone chip with a volume of one million will have a foundry price of more than 150 yuan.

When the quantity reaches five million, the OEM price will drop to around 120 yuan.

When the quantity reaches tens of millions, the OEM price will drop to around 110 yuan.

When the volume reaches 90 million, the OEM price will drop to around yuan.

If the amount is above 70 million, the price will only be around yuan.

Typical economies of scale.

This is also the reason why small companies cannot afford to make chips.

If we make a small number of them, say just one million pieces, the OEM price will be 150 yuan, and the diluted R&D cost will also be frighteningly high. Adding everything up, the cost will exceed 300 yuan, which is more expensive than buying ready-made ones, and the performance may not be that good!

There is no way. These days, 28nm technology is the most advanced technology, but it is expensive.

Once 20nm and 16nm are available, 28nm will become cheaper.

Zhang Chao changed the subject: "But this is our first cooperation. With your 4500 million chip size, we can give you unprecedented discounts. The foundry price of Kirin 920 and 925 within one year can be calculated based on the chip price of Kirin 910."

"Really?" Yu Dong was a little surprised!
The OEM price for a volume of 500 million is 2000 million, and the OEM price for a volume of 2000 million is million. The two are completely different!
For example, for the Kirin 910, which has a volume of 90 million units, TSMC quoted a price of yuan.

后续麒麟920、麒麟925，单款体量也就400-700万颗，那就得120元左右。

As for the Kirin 928, which is expected to produce at most one million units, the OEM price is 150+!

Yu Dong has a headache!
Even if Xingyi Technology gives some discounts, it will only be a few dollars cheaper.

But now, Zhang Chao means that the Kirin 120 and Kirin 920, which originally cost 925 yuan, will be priced according to the Kirin 910, which is only 90 yuan?
This is equivalent to a 30 yuan reduction in the unit price of these two chips, and more than a thousand chips will cost more than million yuan less!
"Is there such a good thing?" Yu Dong was a little bit unbelievable.

“Of course, if your total orders for the three chips reach 4500 million within a year, the subsequent Kirin 920/925 will be priced at the ultra-low price of Kirin 910,” Zhang Chao said with a smile.

"This..." Yu Dong was silent. He just boasted about 4500 million. It was impossible to achieve that. Even for Huawei, not every mobile phone model is a hot seller these days.

On the contrary, it would be great if the sales of mobile phones equipped with Kirin chips could reach 30 million units a year.

"What if there are 30 million..." Yu Dong couldn't help but ask.

"3000 million? That's a bit difficult. After all, the Kunpeng chips purchased by other brands are worth to million, right?" Zhang Chao said meaningfully.

Yu Dong understood what the other party meant and muttered in his heart: "Is this because there are not enough orders for Kirin chips, so I want to use Kunpeng chips to make up for it..."

"But it's not impossible!"

"In the next year, Kirin chips are estimated to account for only 40% of smartphones, and the remaining 60% will have to be purchased from outside."

'Either buy MediaTek or buy Qualcomm...'

'But you can also buy Kunpeng chips from Xingyi Technology, which can also enable Kirin 920 and Kirin 925 to be manufactured at super low prices! '

"TSMC thinks we are small, so the foundry price we offer is not cheap. As for the Kirin 920 and Kirin 925, both are priced based on a volume of million, and the unit price is reduced by yuan, TSMC will never agree to it."

"Xingyi Semiconductor's price is much more reasonable than TSMC, so it is not impossible to use Kunpeng chips."

'And Kunpeng chips are all 64-bit, with strong performance, and 64-bit improvements, which can boost sales! '

Yu Dong quickly made a decision: "That's no problem. As long as the foundry price is low enough, we can purchase another 40 million Kunpeng chips. But the details still need to be discussed with me."

"Of course." Zhang Chao stood up and extended his hand: "I look forward to our long-term strategic cooperation."

"I'm looking forward to it too." Yu Dong said with a smile.

Afterwards, Yu Dong quickly found Ren Dong to discuss the matter.

Director Ren was delighted: "This kind of cooperation is great. We can buy chips from anyone, but if we buy chips from Xingyi Technology, we can also lower the foundry price. Why not?"

"Then we will purchase another 40 million Kunpeng chips?"

“Yes, and the OEM price of Kirin 920 and Kirin 925 is lower, so we can produce more.”

"I understand. In addition, I will ask Xingyi Semiconductor to give us priority in our subsequent tape-outs, so that there will be no problem in long-term cooperation."

Yu Dong added that TSMC is a large company with a huge business, and they have to queue up for wafer production.

They are all Apple, Qualcomm, MediaTek, Nvidia, HiSilicon...

That’s right, NVIDIA is ahead of HiSilicon.

As for MediaTek, its size is much larger than HiSilicon, so HiSilicon can only be ranked last.

For example, Qualcomm did not even get a chance to take part in next year’s 20nm process, as all of it was given to Apple.

The next day, Yu Dong met Zhang Chao again. After a brief discussion, the two smiled:

"Mr. Zhang, happy cooperation!"

"Happy cooperation!"

The two parties have formally reached a strategic cooperation, which is in the interests of both parties.

As for Xingyi Semiconductor's foundry, it will reveal the details of Huawei's chips...

This is not the case. After all, Huawei's current chips are all 32-bit, and Xingyi Semiconductor doesn't even look down on them. They started with 64-bit processors. How could they reverse engineer an outdated 32-bit processor?

This is just a joke.

It's like a doctoral student plagiarizing an elementary school student's composition. It doesn't make sense.

Moreover, Xingyi Technology's chips are more powerful and advanced, and even the future 64-bit processors are not as good as them.

No need to worry at all.

Two days later, the two parties officially signed the contract.

Xingyi Semiconductor gave Huawei a price that it could not refuse. The subsequent prices of Kirin 920 and Kirin 925 will be calculated based on the size of Kirin 910, all at 90 yuan!
In exchange, Huawei purchased 4000 million Kunpeng chips from Xingyi Semiconductor at one time.

其中500万颗鲲鹏706 SOC，1000万颗鲲鹏508 SOC，500万颗鲲鹏508C，2000万颗鲲鹏505 SOC。

并且冲着超低代工价，华为不仅一次性下单了2000万颗麒麟910，连麒麟920的订单量都增加到了800万颗，麒麟925的订单量增加到1000万颗！

If the Kirin 120 or 920 are priced at 925 yuan each, Huawei would not be willing to buy more.

But if you don’t take the opportunity to place more orders for Kirin 90 and 920, which are priced at 925 yuan each, you’d be a fool.

Add to that 100 million Kirin 928s, and a total of 3900 million flagship chip foundry orders have been placed.

Of course, 100 million Kirin 928s will be priced at the normal price of 150 yuan.

The order volume is too small, and if Wang Yi charges 90 yuan, he will lose money.

The purchase of 4000 million Kunpeng chips resulted in ultra-low foundry prices for two flagship chips.

Huawei felt that the remaining amount was more than 4 million yuan in foundry fees, which was well worth it.

Xingyi Semiconductor grabbed a big customer with unlimited potential, defeated TSMC, and sold 4000 million 28nm Kunpeng chips at one time, winning a large foundry order of 3900 million chips!

Wang Yi felt that it was a good deal and he could obviously earn more.

For a big client, it doesn't matter if I earn less in OEM fees.

What’s more, the profit from selling Kunpeng chips has been recovered.

For example, Kirin 920, even with a size of 90 million, Xingyi Semiconductor still makes a lot of money based on the foundry price of yuan.

There is no other reason. The cost of Xingyi Semiconductor's 28nm flagship chip is only about 50 yuan.

Even if the OEM price is 90 yuan, there is still a profit of 40 yuan.

The profit of one chip is 40 yuan, so 3900 million chips would be more than 15.6 billion yuan.

As for the 4000 million chips sold, although they are mainly mid- and low-end chips, the profit per chip is still more than 150.

4000 million is more than 60 billion!

Huawei was given a 4.5 million yuan discount on foundry fees in exchange for 15.6 billion yuan in foundry profits and 60 billion yuan in chip profits.

4.5 million turned into 75.6 billion, a total bargain!

All I can say is that this is a collaboration that both Xingyi Technology and Huawei are very happy about, but only TSMC is hurt!
However, both parties tacitly agreed not to publicize this matter.

From negotiation to signing, everything was done in a low-key manner.

But there is no wall that is impenetrable.

Within a few days, TSMC got the news.

"Xingyi Semiconductor has won Huawei's order?"

Old Zhang opened his mouth wide, a little stunned: "Aren't they short of production capacity? How can they still steal our TSMC customers?"

"Chairman, Xingyi Semiconductor lacks 40-nanometer production capacity, but not 28-nanometer production capacity. So they snatched Huawei's order away."

"Damn it!" Old Zhang was furious: "Wang Yi is here to avenge us for stabbing him in the back!"

"Of course we did. We promised Qualcomm that we would stop supplying Xingyi Technology with 40-nanometer chips. Wang Yi holds a grudge against us and directly snatched Huawei, a major customer, away from us in order to retaliate against us!"

"But it's not Xingyi Technology's fault. If we were suddenly cut off from supply and suffered heavy losses, we would also be angry and seek revenge."

"That's what we said, but we can't swallow this."

"Huawei's orders are not as good as Apple and Qualcomm's, but they are also more than 910 million, especially Kirin 2000, they placed a one-time order of million. Now all of them have been snatched away by Xingyi Technology, and I can't swallow this."

"You can't swallow an order for 2000 million chips a year? If we give up the order from Xingyi Technology, it's 2000 million chips a month! How can you swallow it?"

"That's not a problem. Compared to Xingyi Technology, Qualcomm is obviously more important, right?"

"Even if we were wrong first, we can't just let Xingyi Technology's grabbing of TSMC end like this."

"Okay, then what do you say?"

"We have delivered 3000 million of the last 1000 million chips produced for Xingyi Technology. Isn't there still 2000 million that have not been delivered?"

"You mean, drag it out and not give it?"

"That's definitely not possible. Twenty million chips, even though they are low-end 40-nanometer Kunpeng 500 and Kunpeng 506 chips, cost 25 yuan per chip, so twenty million chips is worth 5 million. If we don't deliver the chips, do we want the million?"

"Not only is the payment of 20 million, but if the goods are not delivered, you will have to pay an additional % penalty, which is another million."

"If we seize their chips, it will indeed add fuel to the fire of Xingyi Technology and cause heavy losses, but it will also cost us 6 million."

"Do you think it's appropriate?"

"Six hundred million in one go! Is it worth it?"

"If it was 6000 million, it would be fine, but 6 million, that would really be..."

"This……"

Everyone was silent.

At this time, the smart Jiang Hong spoke up: "That's not how it works. We did lose 2000 million by deducting 6 million Kunpeng chips, but we got 2000 million Kunpeng chips! We can sell these 2000 million Kunpeng chips!"

"???"

"!!!"

For a moment, everyone was shocked and looked at Jiang Hong with complicated expressions.

Isn’t this method a bit too desperate?

Jiang Hong continued, "Just think about it, Xingyi Technology's Kunpeng 500 is 150 yuan per piece. Kunpeng 506 is 180 yuan per piece. If we sell them in packages, all at 150 yuan, 2000 million pieces will be 30 billion!"

"Lost 6 million, earned 30 billion, net profit exceeded 24 billion!"

"Just tell me if it's worth it!?"

Everyone was speechless: "..."

As expected, senior executives who have been promoted through technical routes are good at technology, but lack a good sense of the overall situation.

(End of this chapter)