Rebirth 99 to become a car giant

Now that Nanshan Semiconductor's Imperial Capital Factory has officially been put into production, it is natural that something has to happen.

Previously, everyone focused on the 14-nanometer process.

This is indeed very important, but for Cao Yang, Nanshan Semiconductor has another trump card that has not been announced to the public, and that is the flagship product of the Imperial Capital Factory.

Compared with the 14nm process chips that many companies are launching one after another, how to produce ultra-high computing power chips based on this chip to provide support for the development of high-end smartphones and electric vehicles is obviously more important.

Under the influence of future cars, the screens of electric vehicles are getting larger and larger, and the number of screens is also increasing.

Models like the W9 not only have a screen in the center console, but there is also a screen in front of the co-pilot.

Behind the front seats, there are also two screens.

In addition, there are many cameras, radars and other parts in the vehicle that need to be controlled by chips, so the computing power requirements for the chips are very high.

That is to say, it was only L2 level autonomous driving before, and the overall requirements were not very high.

So by adding a few more chips, it can basically be done.

But each screen requires its own independent chip to control, and the cost is very high.

And although the computing power of these chips is already relatively powerful, it is just that.

This time, the Nanshan Semiconductor Imperial City Factory is preparing to completely subvert this situation, using the 14-nanometer process to produce ultra-high computing power autonomous driving main control chips, or to produce smart cockpit chips.

The self-driving main control chip is used by myself. It is considered the core competitiveness of future cars. It has no intention of selling it externally, so it does not plan to promote it to customers.

But the smart cockpit chip is different. This can be sold casually and used to make a lot of money.

Compared with previous car-grade chips, the NS8155 chip launched by Nanshan Semiconductor this time was developed to fully benchmark the performance of the later Qualcomm 8155 smart cockpit chip, but the chip manufacturing process has not yet reached 14 nanometers.

But the final computing power reached 8TOPS (1TOPS means that the processor can perform one trillion operations per second).

For semiconductor companies in this era, GOPS (1GOPS means the processor can perform one billion operations per second) and MOPS (1MOPS means the processor can perform one million operations per second) are familiar units.

Nanshan Semiconductor's NS8155 immediately increased the computing power to the order of TOPS. The impact on the entire chip industry is absolutely huge.

"Mr. Cao, Mr. Rao, and Mr. Zeng, our NS8155 is a powerful 8-core automotive chip. It adopts an advanced 1+3+4 8-core design."

"The large core has a main frequency of up to 2.96GHz, three high-performance cores have a main frequency of 2.42GHz, and four low-power small cores have a main frequency of 1.8GHz."

"Such a configuration makes its CPU performance surpass all current automotive-grade chips in the industry."

Zhang Jing personally explained the situation of the NS8155 chip to Cao Yang, Rao Yongxiang, Zeng Tingting and others.

According to the plan, this chip will be installed on the future car S9 model for the first time.

The previous trial vehicle used the NS8155 produced in the laboratory. Now it has entered mass production. Whether it can meet all aspects of performance requirements must be carefully evaluated.

Compared with mobile phone chips or computer chips, the difficulty of developing automotive chips is that they need to adapt to harsher working environments. Whether it is high or low temperature environments, they must be better adaptable.

At the same time, there are clear specifications for corrosion resistance, dustproof and waterproof standards.

This is also to ensure the safety of vehicle driving. Otherwise, it would be too dangerous if the screen, head-up display, and smart cockpit are all unusable when the screen is black while the vehicle is driving.

So normally, the development of automotive chips requires a long process, and the testing process is particularly complex and lengthy.

The development goal of the NS8155 chip is to become the most top-notch configuration in the industry, achieving a good balance in computing power, performance, and software adaptability.

"Academician Zhang, can you briefly explain the benefits of this NS8155 chip?"

Zeng Tingting is in sales, so she naturally hopes to introduce her technical advantages to consumers in simple and easy-to-understand terms.

Otherwise, even people in the automotive industry cannot understand what you are saying, and there will definitely be no way to resonate with consumers.

"Nowadays, the main functions of high-end electric vehicles generally include multiple display screens in the cockpit, maps, Internet of Vehicles, car APPs, 360 panoramic cameras, automatic parking and driving recorders."

"These functions currently give car owners the most intuitive experience whether the car is smooth or not."

“After using our NS8155, you don’t need to worry about the car running smoothly.”

"At least in the next three years, the performance of our NS8155 will be world-leading."

At this moment, Zhang Jing was full of confidence.

Although many chip companies are also optimistic about the development of new energy vehicles and products closely related to electric vehicles such as smart cockpit chips.

Even TI, Qualcomm, Nvidia, Intel, Renesas and other companies have made some arrangements in this regard.

Especially Nvidia, as a former subsidiary of Nanshan Semiconductor, although it was forced to sell out, Nanshan Semiconductor is still very aware of Nvidia's movements.

Everyone is investing R\u0026D resources to develop smart cockpit chips related to electric vehicles and other products that are completely different from existing fuel vehicle chips.

An ordinary fuel vehicle chip has a computing power of 8TOPS, let alone 8TOPS.

Even if you use a 200-nanometer process to produce a chip with 1GOPS computing power, it will be enough for ECU, ESP and other control parts.

After all, the data these control systems need to process is far lower than that of smart cockpit chips.

“How is Qualcomm’s current research on similar products?”

Since the NS8155 draws on many of the layout and performance arrangements of the 8155 chip released by Qualcomm in 2019, Cao Yang must be paying close attention to Qualcomm's news.

"Qualcomm's ambitions are not limited to the mobile phone field. It is also actively deploying in the automotive field, especially smart cockpit chips."

"This is an area that has many similarities with mobile phone chips, but has different challenges and opportunities."

"According to the news we have heard, Qualcomm will officially release the first-generation digital cockpit platform Snapdragon S602A series next year. This is Qualcomm's first step into the automotive field and the beginning of Qualcomm's smart car business."

"S602A is a quad-core processor based on 28nm process, mainly used in car entertainment systems. Our preliminary information is that the performance and quality of this chip are not outstanding. It is estimated that it will not be used in our NS8155 by then. What's the threat?"

"However, they didn't know before that we developed the NS8155 based on the 14-nanometer process, and its performance is countless times higher than that of the S602A."

"After our NS8155 is officially released, I estimate that Qualcomm will soon design a new smart cockpit chip based on the 14nm process."

I have to say that Zhang Jing's vision is still very good.

He can also speculate on some things that have not happened yet based on the current information.

Historically, in 2016, Qualcomm released the second-generation digital cockpit platform Snapdragon 820A series, which was Qualcomm's first major breakthrough in the field of smart cockpits.

The 820A is a quad-core processor based on a 14-nanometer process. Its performance and functions have been significantly improved. It can support more sensors and security, and can also achieve multi-screen interaction and 3D display.

820A has begun to be favored by some luxury car companies, such as Audi, BMW, Mercedes-Benz, etc., and is also adopted by some Chinese new energy car companies, such as Ideal, Xpeng, Lynk \u0026 Co, etc.

Now with the emergence of Nanshan Semiconductor, Qualcomm's 14nm process smart cockpit chips will definitely arrive early.

"What's the situation at NVIDIA?"

Confirming that Qualcomm has no way to threaten Nanshan Semiconductor for the time being, Cao Yang continues to pay attention to Nvidia, a relatively special competitor.

As Nvidia, which uses GPU as its flagship product, has unique advantages in developing smart cockpit chips.

Because a lot of the computing power of the chip in the cockpit is used to process various images, the GPU module integrated inside is needed to function.

When Nanshan Semiconductor acquired Nvidia, it took away a lot of its GPU technology and people.

Otherwise, Nanshan Semiconductor's NS8155 may not be able to be produced so smoothly this time.

"NVIDIA launched the DRIVE platform this year. DRIVE is an end-to-end platform for the development and deployment of autonomous vehicles."

“On the development side, DRIVE includes Replicator synthetic data generation, AI infrastructure, DRIVE Sim and DRIVE Map.”

“In terms of deployment, DRIVE includes full-stack driving and in-car AI applications, AI computers and the Hyperion autonomous vehicle reference architecture.”

When talking about his professional matters, Zhang Jing immediately came up with various professional terms.

Zeng Tingting on the side couldn't help but interrupt: "Academician Zhang, to put it in layman's terms, is NVIDIA's development platform very powerful?"

"Yes, very powerful!"

Zhang Jing responded with certainty, but he immediately added: "When we launch NS8155, we will also launch an industrial software simultaneously to facilitate various suppliers and OEMs to develop NS8155 compatibility on it. "

A chip cannot be directly installed on a circuit board and used. Instead, it requires targeted adaptive development and special software to control these things.

In the traditional fuel era, this part of the development task was generally left to professional auto parts manufacturers such as Bosch, Denso, and Nanshan Auto Parts. The OEMs only had to provide demand.

But entering the era of new energy vehicles, the situation began to slowly change.

On the one hand, there are a lot of things controlled by smart cockpit chips, and the price is also very high. If it is left to suppliers, it will be difficult to reflect the core competitiveness of car companies.

The cost is also not cost-effective.

On the other hand, there are many related parts involved in smart cockpit chips. You can't just leave them all to one supplier to produce, right?

How will you manage such suppliers in the future?

"Our NS8155 is so much ahead of our competitors, it is definitely a booster for the take-off of electric vehicles."

"In the future, I think it will be possible to turn something unfamiliar to ordinary consumers, such as chips, into a core publicity point."

Zeng Tingting is now more confident about the future of automobiles.

When the time comes, the product performance of future cars will definitely beat their peers.

And if you take the lead once, it will be easy to lead step by step.

For example, the future S9 will be the first to use NS8155. Later, other electric vehicles in the industry will use NS8155. As long as this is promoted, it will be difficult to avoid the future S9 car.

Similarly, after Nanshan Semiconductor upgrades NS8155, it will definitely be the first to be installed on future automobile products.

In this way, every future car product can be industry-leading, and the industry's leading position will naturally be established.

This is obviously very beneficial to the development of future automobiles.

"Yes, for electric vehicles, the smart cockpit chip is definitely something to write home about, and it is also a product that consumers can experience the difference in fluency when using the central control screen."

"Don't look at the electric vehicles on the market now. Many of them have made their screens larger, and they have more than one screen."

"But the feedback from consumers is not very good. Various problems such as frequency lag, heating, and black screen often occur."

"Even our current W9 models have more or less problems in this regard."

"Once our NS8155 chip is used, these problems basically disappear."

"Not to mention other things, just fluency can defeat all peers."

As the flagship product of the Imperial Factory, the NS8155 is the result of Zhang Jing’s great efforts. Naturally, he hopes to sell this product well.

"How much does this chip cost, and at what price are you planning to sell it to the outside world?"

Seeing that Zhang Jing had already introduced many of the great features of NS8155, Cao Yang naturally began to care about a very critical issue.

After all, no matter how good the performance of NS8155 is, it will not work if the price is sky-high.

In the previous life, the ex-factory price of Qualcomm 8155 was only US$40. However, in 2020, due to the shortage of overall chips, the price was raised to more than US$200, and even in extreme cases, it exceeded US$500 per chip.

Of course, these are certainly not cost prices and are not very representative.

"According to our theoretical calculations, the cost of NS8155 can be controlled within 100 yuan. If the quantity is large, it is possible to reduce it to 50 or 60 yuan."

"However, the current yield rate has not increased yet, and the cost should be around 200 yuan."

"We can only sell it for 300 yuan when selling to future cars, but for other domestic customers, I think we should still sell it for at least 800 yuan each at this stage, and the price can be gradually reduced to 500 yuan in the future."

Zhang Jing has no intention of selling NS8155 at a low price.

Nanshan Semiconductor's current labor costs are very high, and there is also a lot of investment in equipment.

You have finally produced a flagship product that can be sold everywhere. If you don’t set a higher profit margin, how can you be worthy of yourself?

However, his price proposal is obviously too conservative.

"The computing power of 8TOPS is now far superior to that of its peers, and there is no opponent."

“Since it is a product without competitors, it must look like it has no competitors.”

"The cost of 200 yuan needs to be kept absolutely confidential and no leakage is allowed."

"When sold to other companies within the group, the sales price is set at 1,000 yuan each, and when sold to companies outside the group, the sales price is set at 2,000 yuan. Otherwise, how can the high-end feel of NS8155 be reflected?"

"Of course, if the number of customers increases, it is okay to move closer to 1,000 yuan, but the price cannot be reduced so fast. We must wait until at least a year before considering price reduction."

When Cao Yang said this, Zeng Tingting nodded and added, "It is indeed beneficial to set the price higher."

"There will definitely be more high-end chip upgrades in the future. If the price is set very low at the beginning and the price increases a lot later, it will be difficult for customers to accept it."

"Besides, setting the sales price a little higher can better reflect the high-end feel of NS8155 when promoting it."

Although this will increase the purchase cost of future cars, for high-end models, the difference in purchase cost of one to two thousand yuan will obviously have a very small impact.

But for Nanshan Semiconductor, it is completely different.

If they sell at this price, they can quickly recover the relevant investment costs, and there will be much less pressure later.

"No problem, then I will arrange it in this direction!"

After hearing what Cao Yang and Zeng Tingting said, Zhang Jing was overjoyed.

This result is much better than I expected.

He was originally worried that Zeng Tingting would lower the price of future cars on behalf of him.

After all, according to the current situation, a future S9 will use two NS8155.

This cost difference is still quite considerable.

After internal communication was completed, Nanshan Semiconductor immediately officially announced that it had launched the NS8155 smart cockpit chip.

[NS8155 ultra-high computing power chip, a booster for the take-off of electric vehicles]

Once this slogan was shouted out and some basic parameters of NS8155 were thrown out, the chip industry immediately went into trouble.

Nanshan Semiconductor quietly dropped an atomic bomb into a calm lake.

"Wang Jing, you should immediately find a way to collect more information about Nanshan Semiconductor NS8155. It would be better if you could dig some relevant engineers responsible for the design of this chip."

"As long as someone is willing to come to Qualcomm, annual salary is not a problem, let alone 1 million yuan, even 1 million US dollars is fine."

“We can even help them with green card issues.”

As the president of Qualcomm, Paul is obviously aware of the advancement of NS8155.

Qualcomm's first-generation digital cockpit platform, the Snapdragon S602A series, began research at the beginning of this year and has now been handed over to Samsung for OEM production. It will be officially launched on the market next year.

This is Qualcomm's first big move to simultaneously develop from the mobile phone industry to the automotive industry. Paul is personally watching the development of this project.

But now I find that the S602A developed by myself is a piece of shit in front of the NS8155, and it is not worthy of carrying shoes for others.

This is certainly not a result that Paul can accept.

When will Qualcomm's technology become such a huge gap?

This is simply unimaginable.

Even if there is a gap in the chip technology between everyone, one is 14 nanometers and the other is 28 nanometers, but normally the chip performance should not be that big of a difference.

So Paul wanted to know the specifics.

The development of the chip industry is very fast. If the manufacturing process does not change, it can be updated once a year.

Especially in the smartphone industry, Qualcomm is used to this rhythm.

But now that it has entered the automobile industry for the first time, it has received a huge blow.

"Nanshan Semiconductor has done a very good job of keeping things confidential this time. We didn't hear anything about it beforehand."

"Even though the commissioning ceremony of their factory was so grand, they did not disclose the status of this chip to the outside world."

"Now it seems that they should have successfully started mass production and conducted some basic tests before officially announcing the news."

"There is definitely no problem for us to inquire about some information, but it may not be that easy to find the core personnel who designed this chip."

"Nanshan Semiconductor's salary for engineers is still very competitive in China. Many people majoring in microelectronics are the first choice to enter Nanshan Semiconductor."

"But I can also give it a try. Maybe there are some people who are interested in working in the United States."

Wang Jing didn't dare to take over this task easily, otherwise she would be in trouble if she couldn't complete it.

Of course, she would definitely not keep her words to herself.

After all, the call came from the president who can decide his own position, so he must have the necessary respect for his rights.

"Go boldly and try. As long as you recruit people related to this project, you can at least double their salary, with no upper limit!"

Paul was also unusually generous.

Qualcomm started to develop from the mobile phone industry to the automobile industry. This grand strategy was formulated by Paul himself.

He wants this strategy to succeed more than anyone.

But now we have encountered the NS8155 stumbling block.

According to the current situation, Qualcomm's products do not have any advantages, and there will definitely be no customers to choose them.

And if there are fewer customers for a product, the cost will also increase a lot.

At that time, it is very likely that Qualcomm's S602A will have much worse performance than NS8155, but the cost will be higher.

This situation is naturally not what Paul wants to see.

"I try my best, but I think we can also find some companies to buy NS8155."

“Take the opportunity of purchasing to see if we can learn more about the relevant information.”

Wang Jing proposed a plan that seemed to be more effective.

Then just go to Nanshan Semiconductor to buy NS8155 from a trust, or not a trust, but a company that really needs it.

Although it is not certain whether Wang Jing's plan will have great effect, the cost of implementation is not very high.

So Paul quickly agreed.

Of course, other competitors have naturally thought of the same idea.

Everyone showed their talents and soon obtained samples of NS8155 and some related more specific parameter information.

However, after getting this information, people from various companies became more stressed.

This gap is a bit big.

According to this pace, even in two or three years, we may not be able to catch up immediately.

But maybe in two or three years, Nanshan Semiconductor's more advanced chips will be mass-produced.

Or maybe the chip process has been upgraded from 14 nanometers to 7 nanometers.

Who knows what will happen in the future?