"Could the tape-out of these three chips be carried out at Xingyi Wafer Factory?"
As he said this, Director Qin was in disbelief.
Opposite Mr. Zhao was even more confused.
It is incredible to be able to develop a 40nm system-level chip and integrated baseband by itself, surpassing giants such as Qualcomm.
Have you ever achieved self-production?
Just kidding!
Do you really think that chip technology research and development is just like building a house, just say it?
Wang Yi smiled faintly: "Yes, it was taped out in our Xingyi wafer factory, so the yield rate is not very high. No, it happened to be a chip with insufficient yield rate. Kunpeng 702 lit up three times in a row.
success."
Mr. Zhao: "..."
Director Qin: "..."
The two looked at each other, big eyes and small eyes.
"Xiao Qin, you don't know about this?" Mr. Zhao frowned slightly, he made an oversight!
They thought that Xingyi Technology could only succeed in baseband tape-out, but it turned out that the SOC chip integrating baseband was also successful!
They thought that Xingyi Technology could at best realize the research and development and design of SOC system-level chips, and it turned out that it also achieved mass production!
Completely self-developed and produced!
The problem is that the authorities don’t know anything about such appalling achievements, not even Director Qin, who has direct contact with Wang Yi!
Totally overlooked!
Director Qin was also speechless: "How did I know about such an incredible thing? Damn it, I couldn't even dream of it! Leader, you really don't blame me!"
"Hahaha!" Mr. Zhao also laughed, very happily.
I also feel that Director Qin is really not to blame for this matter. If it were him, he would never have imagined that Xingyi Semiconductor could develop so fast.
But the problem is that it really happened!
It's just incredible.
"Awesome, really awesome, well done!"
Mr. Zhao was very excited and couldn't help but said: "Dr. Wang, what is the yield rate of your 40nm process?"
This is also critical.
Mr. Zhao thinks that if it lights up three times and succeeds once, it is estimated to be only 40 to 50%.
But it’s pretty good, after all, everything is difficult at the beginning.
Wang Yi smiled faintly: "The yield rate of 40 nanometers was 81% before, and we are still waiting for the results this time. The two leaders took a break for tea first, and are currently lighting up all the other chips for testing. They will be ready soon."
It can produce results.”
"No, no need, we're not thirsty."
"Yes, Director Wang, let's watch the lighting test together!"
"Okay!" Wang Yi smiled.
Then I returned to the laboratory with the two leaders.
At the scene, Xingyi Semiconductor workers were conducting tests.
During this tape-out, a total of 1,536 chips of the three types of chips were produced. Once the lighting test is completed, the yield rate of each chip and the overall yield rate will be known!
The surface area of ​​a 12-inch wafer is approximately 70,659 square millimeters, while the size of mobile phone chips is generally around 80-120 square millimeters.
Generally speaking, the area of ​​a chip without integrated baseband is about 80 square millimeters.
The integrated baseband chip is about 100 square millimeters.
Dividing the surface area, a 12-inch wafer can theoretically cut out 700 mobile phone chips.
But obviously, this is not realistic at all. The leftover materials from the wafer cannot be made into chips at all and will be wasted.
It is precisely because of this that the wafers are round instead of square, which would result in more waste.
After deducting the leftovers, only about 640 mobile phone chips can be cut out of a 12-inch wafer.
If the yield rate is 85%, that is 640*0.85=544 pieces.
If the yield rate is 90%, that is 640*0.90=576 pieces!
Of course, this is a system-level chip integrating CPU, GPU, baseband and other modules.
If it's just a separate baseband chip, it will be smaller.
Inside the SOC chip, the baseband only occupies about 10 square millimeters. But a separate baseband chip generally occupies about 30-70 square millimeters.
After all, a separate baseband chip and many modules cannot be shared with the processor, and all must be configured again.
The Kirin 990 with integrated 5G baseband is about 113 square millimeters.
The Kirin 980 without integrated baseband is about 75 square millimeters.
The external 5G baseband Balong 5000 is about 85.83 square millimeters. It is a very special baseband and one of the few basebands that is larger than the chip.
Kirin 980 + Balong 5000 add up to more than 160 square millimeters, which is much higher than the 113 square millimeters of Kirin 990 with integrated baseband.
This is the benefit of integrated baseband, which can save more space for other components.
Of course, Balong 5000 is an exception. Normal baseband chips are mostly 30-70 square millimeters.
Therefore, from the same 12-inch wafer, 1000-1500 baseband chips can be cut out.
The cost of baseband chips is also nearly half cheaper than the cost of SOC chips.
The reason why Qualcomm sells it so expensive is just to earn Qualcomm tax.
Or force users to give up purchasing chips from Nvidia and other friendly vendors and choose Qualcomm's processor chips + baseband chips.
Kunpeng 700 and Kunpeng 702 are both integrated baseband chips, and the area is naturally larger, about 100 square millimeters.
The baseband of Pterosaur 3000 is only 50 square millimeters.
This time, a total of two wafers were used to tape out, producing 512 Kunpeng 700, 512 Kunpeng 702, and 512 Wing Loong 3000.
Under the multi-person test, soon all the chips were lit and the test was completed.
Williams came over with the statistical results: "Chairman, the results are out."
"read."
"Kunpeng 700, 461 were successfully lit, with a yield rate of 90%."
"Kunpeng 702, 435 pieces were successfully lit, with a yield rate of 85%."
"Pterosaur 3000, 471 were successfully lit, with a yield rate of 92%."
“The overall yield rate of Xingyi Wafer Fab’s 40nm process is 89%!”
"So high!" Wang Yi beamed with joy.
Others were also excited.
The yield rate tested last month was only 81%!
This time, it has reached 89%!
I can only say that progress has been rapid.
"The 40-nanometer process is completely done!" Wang Yi looked at Mr. Hu with a smile and gave a thumbs up: "Mr. Hu is mighty!"
"Fortunately, I did not disgrace my life!" Mr. Hu smiled and breathed a sigh of relief.
Wang Yi thinks so highly of him that it would be unjustifiable if he failed.
"Well done! I must celebrate Mr. Hu, Williams, Pang Liguo, and the three teams!" Wang Yi said with a smile: "Tonight, I will be the host of a five-star hotel!"
"Thank you, Chairman." Everyone cheered.
They know that the five-star hotel is just the beginning. This month's performance and year-end bonus will be rewarded.
The team of George and Zhang Chao also became more motivated.
This time the Williams team has successfully developed Kunpeng 700 and Kunpeng 702, and will receive high bonuses.
But George's Kunpeng 901 team and Zhang Chao's Kunpeng 902 team do not have this honor.
Pang Liguo's 3G baseband team successfully developed the Pterosaur 3000 and will also receive high bonuses.
But the 4G baseband team does not have these benefits and has to wait until the 4G baseband is successfully lit!
Wang Yi implements quantitative management. Bonuses are not for one department to eat from a large pot, but are accurate to specific project teams and individuals.
A department, a different project team, and different breakthroughs will all receive very different treatment.
"Mr. Hu, I'll leave the mass production to you. Let's produce Kunpeng 702 first!" Wang Yi instructed.
"Okay, Chairman, I will make arrangements right away!" Mr. Hu responded, and then said meaningfully: "Director Wang, don't forget the new wafer fab."
"Don't worry." Wang Yi said, looking at Director Qin and Mr. Zhao: "Both leaders, let's talk inside."
"good!"
The three of them returned to the office, and Mr. Zhao could no longer suppress his excitement:
"Dr. Wang, this is amazing! It's amazing! Not only has the 40-nanometer chip been mass-produced, but it has also achieved a yield rate of 89%, awesome!"
Wang Yi smiled lightly: "Actually, it's more than 89%. The 11% of scrap chips will also be further processed to make a low-frequency residual version of Kunpeng 700!"
"I see. In this way, the scrap rate is actually lower." The two of them suddenly realized.
"Yes, two leaders, next we will give priority to mass production of Kunpeng 702, and replace the Xphone 1pro, Xingyi Tablet X1pro, Xingyi TV X40, X50 and NVIDIA Tegra 3 with self-developed and self-produced Kunpeng 702. Also
The APQ 8064 of xphone 2 has also been replaced with the self-developed and self-produced Kunpeng 702!"
"Great!" Mr. Zhao was excited: "You are the first company in mainland China to realize self-development and production of 40-nanometer chips, and also the first company in the country to realize self-development and production of 40-nanometer chips."
"The pride of the country, the pride of the nation!"
"Dr. Wang, you did a great job."
"Tomorrow I will ask CCTV to come to Xingyi Technology and Xingyi Semiconductor for an exclusive interview, which will be broadcast on CCTV News and serve as a technological benchmark and an example for all companies!"
Wang Yi waved his hand quickly: "Mr. Zhao, there's no need for that. At this stage, it's better to hide one's strength and bide one's time and walk at night in brocade clothes."
"This..." Mr. Zhao frowned slightly: "You are worried..."
Wang Yi sighed: "Mr. Zhao, although Xingyi Technology has achieved self-development and production of 40-nanometer chips, core equipment such as photolithography machines and core raw materials such as photoresist cannot be produced domestically, and they all rely on imports."
"At this stage, it is okay for us to break the blockade, realize self-research and production, and make a fortune silently. If we are too high-profile and cause unnecessary trouble, once we are stuck, on the contrary..."
"Yes, what you said makes sense." Mr. Zhao nodded, his expression a little complicated.
A little frustrated and helpless.
But the situation is stronger than the person, so I have to bow my head.
Wang Yi continued: "And I plan to not make any announcement in these two months, secretly mass-produce Kunpeng 702, and carry out adaptation and training. After two months, I will directly replace it and launch it directly! Competitors such as Apple, Qualcomm, and Samsung will be replaced.
Caught off guard!"
"If it is exposed now, if they secretly plan to prevent the mass production of Kunpeng chips, we will still be the unlucky ones."
For long-term interests, Wang Yi does not seek temporary gains and losses.
"Yes, your consideration makes sense. Don't worry, you have made breakthroughs in 40nm process, and the official will pay more attention to your previous proposal. Relevant institutions have been established and a team has been established to further increase the inspection of photolithography machines and photoresist.
R&D investment in other fields should be regarded as the core project to achieve breakthroughs."
Mr. Zhao sighed: "In the past, various factors determined that our investment in semiconductors and lithography machines was not large enough. But now, it has completely changed."
Initially, the report Wang Yi asked Director Qin to send attracted some attention, but it was very limited and was quickly shelved.
Later, when Xingyi's mobile phones were selling like hotcakes, Director Qin went to the old leader Zhao again with the report and gave him earnest advice.
Mr. Zhao also felt that Xingyi Technology had a future and once again presented the report at the top-level meeting.
This time, it has the personal endorsement of ministerial leaders, and Xingyi Technology continues to grow, becoming a top giant.
I can only say that a person’s name is like the shadow of a tree.
The previous Xingyi Technology was an emerging company, and the upper management did not pay enough attention to it. No matter how good the report was, they did not pay much attention to it.
But now Xingyi Technology has become the top domestic giant and is invincible internationally. It can defeat Apple and Samsung, and HTC can't be beaten.
It can be said that things are no longer what they used to be!
There was no surprise. This time, Wang Yi's report attracted great attention from the top and received full support from many dignitaries and academicians.
Projects such as photolithography machines and photoresists have also been successfully included in the national strategic core projects, requiring relevant agencies to jointly tackle the problem and invest huge sums of money to achieve breakthroughs.
Now that Xingyi Semiconductor has achieved self-development and production of 40-nanometer chips, Wang Yi's weight has become even heavier.
The gold content of that report will be higher, and the research and development and investment in related projects will be even more terrifying!
To put it bluntly, strength speaks for itself in everything.
In your previous life, why did you support Huawei so much?
It is because Huawei has the foundation, ability, and technology to achieve breakthroughs that other companies cannot.
Similarly, Xingyi Technology also showed terrifying strength.
Xingyi Semiconductor has achieved baseband integration that Apple, Samsung and Qualcomm cannot do.
The 40nm process that SMIC has been unable to master for so many years has been mastered by Xingyi Wafer Fab. The trial production yield rate is as high as 89%, and the follow-up will only be higher.
What Huawei cannot do is develop and produce chips by itself. Xingyi Technology has done it with excellent performance.
Such an achievement will shock the whole world.
The senior management will naturally pay close attention to it and provide strong support.
"Dr. Wang, don't worry. In a few years, photolithography opportunities will break through, and photoresist will also break through!" Mr. Zhao said solemnly.
The lithography machines these days are still DUV lithography machines. They are not as difficult as EUV lithography machines, so it is naturally easy to make breakthroughs.
Similarly, photolithography machines only produce mid-range photoresists of 28 nanometers and 20 nanometers, which are not as difficult as the high-end photoresists of 5 nanometers and 3 nanometers in the previous generation.
To put it bluntly, the current gap with Europe and the United States is still very small.
Because Wang Yi had no idea of ​​the high-end proposal before, the official made plans in advance and tackled the tough issues in advance. When the gap was small, it was possible to catch up with Europe and the United States in advance, and even overtake in corners.
In my previous life, I was too far behind.
At that time, European and American EUV had been updated for several generations and had been in use for many years. It was naturally difficult for us to catch up with DUV lithography machines.
But this time, everyone is still DUV, so it will be simple.
As for ASML’s EUV, it is not mature yet and will take several years to be officially put into use.
Because of Wang Yi's suggestion, if you catch up in advance, you will have a bright future.
"That's perfect. I believe in the power of the state machine." Wang Yixiao said.
For example, it will be difficult for Xingyi Technology to build a DUV lithography machine in ten years.
If the government wants to do it, and with the full support of the state machinery, 14nm DUV will be almost ready in five years.
Subsequent iterations and upgrades will continue to gradually improve the accuracy, reaching 10 nanometers and 7 nanometers.
After all, ASML launched the first-generation immersed DUV 1250i in 2003, and initially manufactured 45-nanometer chips.
Subsequently, we will continue to carry out comprehensive upgrades on the light source, optical system, overlay accuracy, temperature control and other aspects of the DUV lithography machine.
Only then did we gradually realize the 32-28nm process, and the subsequent 20-16nm process, and then the 10nm node, the 7nm node, and the 5nm node ten years later.
The main reason why domestically produced DUV lithography machines have not been mass-produced in the past is that the requirements are high and the starting point is high.
If we can only mass-produce 45 nanometers, it won’t make much sense, and not many companies will use it once it’s launched.
Ten years from now, what will be lacking is high-end lithography machines, not 45nm lithography machines.
But if you want to mass-produce 7 nanometers or even 5 nanometers, it will be very difficult, and domestic DUVs will have to be continuously upgraded.
Therefore, the delay in launching domestic DUV is just holding back the big move.
But this time it doesn't matter.
If you start early and can produce 28 nanometers, you will be considered very successful.
Even if 45nm is mass-produced, Wang Yi will buy it like crazy.
After all, 45nm will be the mainstream chip in the next five years.
28nm and 20nm are only a small number of flagship chips, accounting for only a minority.
Instead, more chips are affordable and high-quality 45nm.
If domestic lithography machines can produce 45 nanometers, there will be a huge market for them.
But it will be difficult if 45-nanometer domestic lithography machines are launched ten years later.
After all, 45 nanometers ten years from now will really be unusable.
Zhao Lao changed the topic: "But you keep a low profile, but you should be rewarded a little more."
"From now on, all Xingyi Semiconductor's chips will be tax-free."
"Xingyi Semiconductor will be tax-free for the first five years before it makes profits!"
"In addition, Xiao Qin made a bet with you before. If Xingyi Semiconductor can successfully mass-produce 40nm and the yield rate exceeds 85% before the end of 2014, it will provide a 12-inch wafer fab."
"If you realize it in advance, we will also pay in advance, and we will give you a super-large fab with a monthly output of 100,000 wafers!"
"Thank you, leader." Wang Yi was in a good mood. This was what he needed most.
It is also what Mr. Hu is looking forward to most.
The current wafer fab has been fully delivered, but its monthly production capacity is only 25,000 wafers.
One wafer can cut out 640 Kunpeng 702 chips. Calculated based on a 90% yield rate, there are 576 high-quality chips.
That’s 14.4 million Kunpeng 702 pills per month.
If used with a single mobile phone, it is completely sufficient.
But the problem is, xphone 2, xphone 1pro, Xingyi tablet X1pro, Xingyi TV X40, x50 all need to use it.
That’s simply not enough!
At present, the monthly sales of Xphone 2 are in the tens of millions, and including parallel imports, 15 million is not enough.
This is difficult to correct.
It can be said that Xingyi Semiconductor now has chefs, but there are not enough stoves and not much cooking can be done!
Director Qin said with a smile: "The site for the new wafer fab has been planned for a long time. Now you are mass-producing 40 nanometers. Go through the procedures as soon as possible and start construction next month! The official support is strong and we will go all out to build it and strive to deliver it all within two years. Of course.
, you have to handle the subsequent production line equipment yourself."
The government is only responsible for investment in infrastructure and giving away land for free.
"Within two years? Including the introduction of the production line and putting it into production, it will take two and a half three years. It is too slow!"
Wang Yi frowned.
"What? What difficulties does Mr. Wang have?" Mr. Zhao was keenly aware of Wang Yi's expression.
Wang Yi sighed: "Boss, I'll just tell you something."
"But it doesn't matter!"
"It's like this, leader. Although the current Xingyi wafer fab has achieved mass production of the 40nm process and has a yield rate of 89%, its production capacity is too small. It has a production capacity of 25,000 pieces per month and can produce a maximum of 14.4 million Kunpeng 702
.”
"This output sounds like a lot, right?"
"But in fact, the xphone 2 alone is not enough!"
"Currently, the sales volume of Xphone 2 is more than 15 million units, and the sales volume of Xphone 1pro is also around 10 million units. We want to achieve full replacement of self-developed and self-produced chips, but it is not enough!"
"Not to mention, there are Xingyi Tablet X1pro, Xingyi TV X40 and X50, which also need Kunpeng 702."
"Next year, the thousand-yuan machine will be mass-produced and will be equipped with the Kunpeng 700. The Kunpeng 700 will also have to be mass-produced, and the annual output will be 100 million."
"Coupled with the Pterosaur 3000 baseband, we can also mass-produce some for external sale."
"Added all together, Xingyi Technology's chip demand will exceed 300 million pieces in the next year."
"The current Xingyi wafer fab can produce up to 170 million wafers a year, and there is still more than 150 million vacancies in production capacity."
"this……"
Mr. Zhao and Director Qin were both stunned.
I had no idea that Wang Yi’s ambitions were so big, nor that Xingyi Technology’s chip demand was so high.
But it’s normal to think about it.
Next year's quad-core Unbounded smartphone will start at 999, and it will definitely sell out. If there is hope of sales exceeding 100 million, 100 million Kunpeng 700 chips will be needed.
Xingyi's new flagship mobile phone, plus this year's Xphone 1pro and Xphone 2, the three models combined have sold over 100 million without any pressure.
Xingyi Tablet is now on sale globally, with old models and new models. Sales will exceed 60 million next year, so there is no pressure.
Xingyi TV’s many new and old models combined can produce annual sales of 50 million without any pressure.
In addition, the 3G baseband Wing Loong 3000 can also be sold to the outside world. Customers such as Huawei can purchase tens of millions a year without any pressure.
It is possible to add up to 350 million.
The production capacity of Xingyi Wafer Fab is only 172.8 million chips, which is simply not enough!
Just half the difference!
As for the second largest wafer fab, it has a monthly output of 100,000 wafers and an annual output of 1.2 million wafers. If it is used to produce Kunpeng 702, the annual output will be 691.2 million wafers.
Coupled with Xingyi Wafer Factory 1, the total production capacity will reach 864 million units.
By then, the production capacity will be sufficient, or even not used at all, and we can still outsource manufacturing.
But that will be two or three years from now, but right now, even in the next year, it won’t be enough.
It’s simply not enough!
Distant water cannot quench near thirst.
Mr. Zhao fell into deep thought: "Director Wang, you are right. What Xingyi Technology needs now is not the super large fab that will be built in two or three years, but a small fab that can be put into production quickly."
"Therefore, I decided to split the second fab with a monthly output of 100,000 wafers into several small fabs to improve delivery time."
A very large fab with a monthly output of 100,000 wafers will take two years to complete and two and a half to three years to put into production.
But if it is divided into small fabs with a monthly output of 20,000 to 30,000 wafers, it will be easy. They can be delivered in about half a year and can be put into production within a year.
"But this is still too slow." Director Qin, who had been silent, suddenly said:
"Mr. Zhao, isn't SMIC investing in and building too many wafer fabs, making ends meet, and losing money every year? We can just pay for it, buy one, and transfer it to Xingyi Technology."
In the past few years, in order to please the local government, SMIC expanded wildly and built factories.
A 12-inch wafer fab was built in the imperial capital. SMIC’s headquarters in Shanghai was not happy, so Lao Zhang built a new 12-inch fab in the capital!
Now that there are both the Magic City and the Imperial City, Shenzhen City must also be built, right?
The result is that SMIC is building more and more wafer fabs. The production capacity of the 8-inch line has not been fully reached, and the 12-inch line has been put into production...
Imperial 12-inch wafer fab, hosting Seoul 12-inch wafer fab, Modu 12-inch wafer fab, Shenzhen...
There are many typical factories with poor students.
It can't be used, it can't be used at all!
Chapter completed!