Rebirth 08: Rise from copycat phones

Chapter 280: Graphics card release and acquisition of Nvidia

In order to make money from mature technology, Zhiyun Microelectronics has built two advanced wafer fabs, No. 17 and No. 18, which focus on nodes below 18-nanometer process.

We also continue to upgrade and transform the technology of several existing wafer fabs, mainly to upgrade the previous wafer fabs with the lowest node of only 32 nanometers to 28 nanometers.

Including a 12-inch factory in Shenzhen and a 12-inch factory in Shanghai.

It is worth mentioning that Zhiyun Microelectronics' 28-nanometer process is itself a slight upgrade of its self-developed 32-nanometer process. The overall technology has not changed much, there is a great deal of technical commonality, and the equipment is the same, so it is also very convenient to upgrade.

This is also the main factor that enables Zhiyun Microelectronics to transform these 32-nanometer process factories into 28-nanometer process nodes after a simple upgrade.

As a result, Zhiyun Microelectronics now has two more 12-inch factories capable of producing products with 28-meter to 45-nanometer processes.

However, not all wafer fabs can carry out this kind of technology upgrade. Many old 12-inch factories and old 8-inch factories cannot upgrade to the 28-nanometer process because a large number of old equipment in the factories cannot meet the technical requirements of the new process.

These old factories can only continue to produce some backward chips with 65-nanometer process, 90-nanometer process, or even some with process above 130-nanometer process.

As for replacing equipment in the old factory... the main investment in the wafer factory, a large number of core equipment such as lithography machines and etching machines are a bulk of the cost. It would be better to build a new factory than simply replacing the equipment.

Now and in the next few years, Zhiyun Microelectronics' logic chip foundry layout will mainly continue to focus on 18nm and 14nm, and look forward to more advanced equivalent 10nm or even 7nm processes.

At the same time, it will continue production and even expand mature 28-nanometer and 45-nanometer production lines. In particular, the 28-nanometer production line is believed to have a large market demand for this process node even many years later.

Because Zhiyun Microelectronics not only invested in building a new generation of wafer fab with more advanced technology, but also invested 2 billion US dollars to start construction of a new 12-inch wafer fab in Ningcheng, focusing on the 28-nanometer process.

However, the backward 65-nanometer process and 90-nanometer process nodes, as well as the even more backward process nodes above 130-nanometer, are basically produced by some existing old factories. These old factories generally do not undergo major technological upgrades and will continue to produce until they are completely outdated and have no orders.

Don’t underestimate these mature and backward technologies. In fact, they can also make money. Because Zhiyun has long recovered the investment cost of the year by relying on a large number of orders and production capacity, the subsequent profits are all pure profits...

It’s just that Zhiyun will not make additional investments in these backward technologies. After all, there are a large number of domestic chip factories catching up in this area.

In recent years, with the rapid development of the domestic electronic consumer industry chain, the demand for various chips is huge. Many people have seen the opportunity to enter the chip design industry. Some companies have also entered the chip foundry industry, or designed their own chips and built their own chip factories for production.

This has led to huge new investments in domestic chip factories, and many small and medium-sized chip factories have emerged, and they mainly produce these medium and low-end processes.

Of course, this is also related to the vigorous promotion of Gulf Technology!

In order to promote its own lithography machines, etching machines and a large number of other key semiconductor equipment products, Gulf Technology has been deceiving others all year round to invest in building chip factories. It also offers preferential prices for various core equipment and even helps you introduce banks to provide loans and other services.

If you don't know how to do it, Gulf Technology can help you build a factory, train technical personnel, and provide you with technical support in the production process.

We can even help you build a wafer factory directly, and then coordinate with Zhiyun Microelectronics to obtain technical authorization for some backward processes.

All you have to do is pay the money and you can get a very high-end wafer factory.

It has greatly lowered the market entry threshold.

Therefore, we can see that many domestic chip factories have emerged recently. These factories all use core equipment such as lithography machines and etching machines from Gulf Technology. The entire semiconductor equipment and consumables supply chain are all under the Fairy Mountain Holdings.

These newly emerged domestic chip factories mainly focus on the 65-nanometer and 90-nanometer processes, as well as the more backward 130-nanometer, 150-nanometer, and 180-nanometer process nodes.

As for more advanced process nodes, such as 40 or 50 nanometers, the investment is too huge and the technical requirements are too high. At least don't expect to easily obtain the corresponding technology license.

These are the people who make a living for several chip giants, so it is impossible for them to easily authorize others.

Therefore, in the domestic chip manufacturing industry, except for Zhiyun Microelectronics and SMIC, there are currently no domestic manufacturers entering the process nodes below 45 nanometers.

However, some domestic chip manufacturers are also waiting for Gulf Technology's DUV immersion lithography machine to achieve a breakthrough...

Although he doesn't have much money and his skills are not that good, he is ambitious... It may not work out now, but who can say what will happen in the future.

------

With the rapid development of the semiconductor industry led by Zhiyun, including chips and manufacturing industries, some top-notch semiconductor products have also been produced.

In late October, Zhiyun Semiconductor, a subsidiary of Zhiyun Group, held a new product launch conference, at which it unveiled a number of chip products.

A new generation of low-voltage DDR3 1600 memory chips using nm process technology.

This type of low-voltage memory is mainly used to supply mobile devices, such as laptops. Its biggest feature is that it greatly reduces power consumption, which is more conducive to extending the battery life of laptops.

At the same time, the company also announced the third-generation 3D NAND flash memory particles and the corresponding flash memory controller, which further increased the number of layers. At the press conference, it also launched the "Zhiyun" brand of consumer-grade solid-state hard drive products, and also released a new generation of Jiangcheng storage brand's cost-effective solid-state hard drive products, launching a new round of impact on the solid-state hard drive market in the PC market.

Among them, the new generation of cost-effective solid-state hard drives of the Jiangcheng storage brand are still quite attractive in price. The price of 64G is 128 yuan, and the price of G has dropped to yuan.

Jiangcheng Storage's solid-state drives, which focus on cost-effectiveness, have always been cheaper than Zhiyun brand solid-state drives. Zhiyun brand's new generation of 64GB solid-state drives this year cost 128 yuan, and GB costs more than yuan.

Of course, although the above two seem to have the same capacity, in fact, the performance gap is very large... The gap in key data such as data reading and writing is very large.

In addition to releasing conventional consumer-grade flash memory products, namely solid-state drives, Zhiyun Semiconductor also announced a new generation of enterprise-level flash memory solutions SF3500 at the press conference.

This large-capacity, high-speed flash memory array, specially provided for enterprises, uses the latest third-generation flash memory particles from Zhiyun Semiconductor and is equipped with the latest developed enterprise-level flash memory controller.

Of course, the price is also extremely high, but even though the price is high, there are still many companies purchasing it on a large scale, many of which are world-renowned companies, such as Zhiyun itself, some domestic Internet companies, and some well-known foreign high-tech companies, such as Google and Microsoft, which are all customers of Zhiyun Semiconductor's SF brand enterprise-level flash memory products.

Zhiyun Semiconductor is currently one of the world's leading memory and flash memory manufacturers, with several brands under its umbrella.

The main ones are Zhiyun, SandForce of the United States, Qimonda of Germany, and Jiangcheng Storage.

Among them, the Zhiyun brand mainly provides memory and flash memory particles, and its main customers are some memory manufacturers, solid-state drive manufacturers, mobile phone manufacturers, etc.

At the same time, we also provide memory and flash memory finished products to the market.

The Zhiyun brand is the most familiar brand of Zhiyun memory and flash memory products in everyone's eyes, and it is also the core storage brand of Zhiyun Semiconductor.

All memory and flash memory chips produced by Zhiyun will have the word YUN on them.

The American SandForce brand, this subsidiary mainly produces enterprise-level high-speed and large-capacity flash memory products. Because its flash memory controller technology is quite outstanding, its customers are all over the world. Google, Microsoft, Apple, Zhiyun and some other high-tech companies are all customers of the subsidiary's SF series products.

Then there is Germany's Qimonda, which is a finished memory brand mainly targeting overseas markets, especially Europe, the United States, India and other markets. Many American and European smart device manufacturers and PC manufacturers have adopted some Qimonda brand memory.

This year, Zhiyun Semiconductor launched a new brand of Jiangcheng storage. The brand mainly targets the domestic market with a cost-effective approach and provides finished memory sticks and solid-state drives to the outside world. Because they use old technology, the prices are relatively cheap.

The above memory and flash memory products are what Zhiyun Semiconductor has been playing with for several years. The market has gradually matured, people are familiar with them, and they are widely supplied to the global market.

In addition to Zhiyun and Weiku Electronics, which directly use it, a large number of smart terminal products and PC products also use Zhiyun Semiconductor's memory and flash memory.

In addition, many foreign manufacturers also adopt it in large quantities.

Zhiyun Semiconductor's position as the world's third largest memory manufacturer and second largest flash memory manufacturer is not just boasting, but is the result of real market experience. Its shipments are enormous.

Moreover, Zhiyun brand memory and solid-state hard drive products are also top-tier high-end products in the domestic DIY market, comparable to four-star products.

Especially for solid-state drive products, currently only Sixin and Zhiyun are playing with 3D NAND flash memory particles on a large scale, competing with each other and continuously stacking layers.

Therefore, in this year's press conference, some new products released by Zhiyun Semiconductor also attracted a lot of attention, but the most popular products at the press conference were not these memory and flash memory products.

It is a field that Zhiyun Group has never been involved in.

Consumer-grade high-performance independent graphics card!

At this year's Zhiyun Semiconductor press conference, Zhiyun Semiconductor officially launched to the market the high-performance independent graphics card with the long-polished OPA core architecture: OPA1000.

The so-called OPA architecture is actually a super modified version of the OP integrated graphics architecture previously used within Zhiyun Semiconductor.

Zhiyun Group's graphics processor technology route is very clear. At the beginning, it acquired the permanent buyout authorization of a foreign mobile graphics processor architecture. Subsequently, the graphics processing project team under Zhiyun Semiconductor began to develop its own graphics processing core based on this foreign graphics processing core architecture.

Subsequently, the AP60 architecture for mobile terminals was launched, and then further developed and surpassed the original, launching the new self-developed AP70 architecture, and then developed the AP75 architecture.

Currently, the graphics processing core of this AP architecture is used in almost all Zhiyun and Weiku brand smart terminals.

At the same time, most domestic smart terminals and some foreign manufacturers' smart terminal manufacturers...as long as they are smart terminals using Zhiyun Semiconductor W series SOC, all use this AP architecture.

Because the graphics processing core in this type of mobile smart terminal is often directly integrated into the SOC. If you want to use Zhiyun Semiconductor's W series SOC, then you can only use the GPU core of this AP...you have no choice.

Because S series chips and W series chips are widely installed in smart terminals of various manufacturers and sold to the global market, this AP architecture has become the largest mobile graphics processing core in the world, with a market share of more than 50%.

Quite exaggerated!
This has led to the fact that most software manufacturers, especially mobile game manufacturers, must actively adapt to Zhiyun's AP graphics processing architecture when developing software, especially games.

Otherwise, it would be tantamount to giving up more than half of the smart terminal market in vain.

If NVIDIA is the graphics card leader in the PC sector, then Zhiyun is actually the graphics card leader in the mobile sector, without a doubt.

Taking advantage of the strength of the AP architecture, Zhiyun Semiconductor modified the AP architecture using this architecture. However, due to its low power consumption characteristics, its performance is not very good when used on the PC side, and it can only be used as an integrated graphics card in the CPU.

It is currently mainly used in Pangu computer CPUs with the SOP instruction set. In the future, it will also be used as the built-in graphics processing core of Zhiyun X86 chips and even future Zhiyun desktop ARM chips.

The AP architecture and the OP architecture are actually very similar. To some extent, the OP architecture is just an amplified version of the AP architecture. The two are highly compatible.

Zhiyun Semiconductor also tried to use the OP architecture to design high-performance independent graphics cards, but the results were not very good... OP continued the low-power consumption characteristics of the AP architecture, but if the architecture focuses too much on low power consumption, this performance cannot be achieved.

Later, the OP architecture was deeply modified, and the low-power consumption characteristics brought by the mobile terminal were abandoned. While continuing to be compatible with AP and OP architectures, a large number of technologies on AI chips were also adopted.

Finally, an OPA framework was developed.

As for why we refuse to give up this low-power AP/OP architecture...it’s naturally because of the ecosystem.

This ecosystem does not simply refer to game support performance, but refers to a complete set of graphics development software systems released free of charge by Zhiyun Semiconductor: CYAP.

This complete set of Zhiyun graphics software development system, which contains a variety of developer software, allows individual developers and software manufacturers to optimize their own products to better adapt to the AP/OP architecture.

This CYAP graphics processing ecosystem is an ecosystem in the field of graphics processing that Zhiyun Group has accumulated with great difficulty.

The vast majority of mobile software on the market, whether it is office software or games, all support or are specially optimized for this system.

It also means that there are a large number of software engineers in the market who have been exposed to or are good at CYAP development.

What Zhiyun wants now is to expand the entire CYAP ecosystem from the mobile side to the PC side and complete the layout of the entire industrial chain of the ecosystem.

If we don't consider the ecosystem, in fact, with the strength of Zhiyun Semiconductor, it is not that troublesome to make the so-called graphics card. Just modify its own AI chip and reduce its size, and it will be almost the same. The hardware is definitely first-class.

But this is obviously not the way to go. The products produced in this way will have poor adaptability to the market ecosystem... If no one supports your ecological content, no matter how powerful your graphics card is, you will not be able to bring it into play.

Therefore, the main focus of Zhiyun Semiconductor's civilian graphics card project is still on the AP/OP architecture, integrating its own unique AI chips and being compatible with some development software for AI chips.

We have created hardware products that are compatible with both the CYAP ecosystem and the AI ​​ecosystem.

Zhiyun Semiconductor wants to integrate all things in the field of graphics processing into one place to create a highly adaptable ecosystem.

Zhiyun Semiconductor has very big ambitions.

The product that integrates AP/OP with AI chip architecture is the OPA graphics processor architecture officially released by Zhiyun Semiconductor today.

Based on this architecture, the first generation of OPA1000 graphics processor chip was produced.

Then, based on this chip, the first generation of OPA1000 high-performance independent graphics card was made.

At the press conference, the people from Zhiyun Semiconductor were quite boastful, saying that the core of this OPA1000 independent graphics card was manufactured using the top-level -nanometer process technology and adopted the core technology of Zhiyun's many graphics processors.

It perfectly supports Zhiyun CYAP technical standards, as well as Microsoft's DirectX 11 and a series of common graphics processing technologies or standards on the market...but many exclusive technical standards of NVIDIA or ATI cannot be used.

At the same time, the hardware has more than 5 stream processing units, uses a GDDR3 video memory interface, has 6000GB of video memory, a current memory frequency of 384MHz, and a video memory bit width of bits, which are performance indicators that are very scary at first glance.

Then he also said that the power consumption of his own graphics card is very well controlled, and the design power is much lower than that of competitors of the same level.

In short, it's all exaggeration. If you really go by these performance data, it can be compared with NVIDIA's latest sub-flagship graphics card.

Even better... because the OPA1000 consumes less power and is more energy-efficient.

In fact, it cannot be said that the exaggeration is wrong, because if we really look at it from a pure hardware level, Zhiyun Semiconductor's chip design capabilities are actually quite excellent, and it is not a problem to make a graphics card that only focuses on computing power.

They can even develop higher-level AI chips, let alone consumer-grade graphics cards.

Although it was boasted a lot at the press conference, if you actually use this graphics card to install a computer and then run games... I'm afraid it will be beaten to a pulp by the flagship N or A cards!
There is no way. The hardware level is sufficient, but the ecosystem is not good.

A large number of games are not optimized for your graphics card, and are not even designed according to your ecosystem. What else can you expect? It's good enough to be able to open the game smoothly and let you play it. Even so, it still depends on the powerful hardware computing power of the OPA graphics card.

Don't expect any optimization.

So those ordinary game players shook their heads after seeing this graphics card… A graphics card with this level of hardware can’t be cheap. It must cost at least three or four thousand yuan. But if you buy a graphics card for three or four thousand yuan, you can only use it as a mid-range graphics card. Isn’t this a big mistake?
Later generations of domestic graphics cards also encountered this embarrassing situation... In fact, the hardware was already acceptable, but due to ecological problems, the gaming experience was relatively poor, even worse than foreign mid- and low-end graphics cards, resulting in poor sales.

However, after Zhiyun Semiconductor released the OPA1000 graphics card, ordinary consumers shook their heads after seeing it, but another unique group of graphics card consumers, that is, a large number of mining bosses, their eyes lit up after seeing this graphics card.

They found that this thing seemed to be specially built for them, with explosive computing power and not too high power consumption, and some of the accompanying ecological development software was particularly suitable for running mining operations.

This is not an illusion, but the graphics card has already integrated some technologies of AI chips. As a gaming card, it may not be very useful due to ecological restrictions, but it is very useful if you use it to run complex calculations, especially in pure computing scenarios such as mining.

This also led to the fact that although Zhiyun Semiconductor's goal was only to create a high-performance consumer-grade independent graphics card, after it was actually produced, few gamers were interested in it. Instead, mining bosses drooled at the sight of this graphics card.

The mining bosses buy graphics cards from the perspective of doing business. As long as the performance is high enough, the power consumption is low, and the overall cost-effectiveness is high enough, they will buy graphics cards even if they cost tens of thousands, let alone four or five thousand yuan.

As a result, the mine owner placed orders frantically...

When mining bosses began to notice this graphics card called OPA1000, which is actually a top-level computing power card, another group also noticed it.

These are people who need to process large amounts of graphic data, such as video editors.

They also had very high requirements for the performance of graphics cards, and then they discovered that... the graphics card produced by Zhiyun was particularly useful for processing video data, and was even better than those high-end graphics cards that traditionally focused on gaming performance.

In their words, using this card to play games is a waste, and using this graphics card to process video data is its best use...

At the same time, some mobile software developers have also discovered that this graphics card is quite special, that is, it supports the CYAP ecosystem and can be used to develop corresponding mobile software, especially mobile game products, which can easily carry out various development and testing processes.

Therefore, after Zhiyun Semiconductor released this OPA1000 graphics card, the response among gamers was average, but it aroused quite a stir among mining bosses and office workers.

This led to an even more interesting phenomenon, which was that after this graphics card was released, few gamers bought it. However, after it was put on several online shopping platforms, especially the Jingxi platform, its sales volume was surprisingly huge.

In November, the OPA1000 graphics card was among the top five in the high-end graphics card sales rankings in Jingxi.

For a newly released graphics card with no large-scale gaming ecosystem or user base, and just an attractive Zhiyun name, this result surprised many people.

Some people who didn't understand the situation continued to admire it, saying that it was worthy of Zhiyun and that anything it bought could set off the market, even a new field like graphics cards.

There are also some half-baked so-called technology bloggers who seriously analyzed why the sales volume of this graphics card is so large, saying that this is due to Zhiyun's brand appeal, coupled with the influence accumulated by Zhiyun in the field of mobile graphics processors over many years, and so on.

But not many people can really hit the point.

It was not until a mine owner posted a corresponding post on an online forum that he said that after using the OPA1000 graphics card, the overall mining income increased by about % compared to using graphics cards of the other two brands. Several other mine owners also agreed with this situation in their replies.

Then someone moved this post from this professional forum to the outside world, and gave it a particularly scary title: "The Zhiyun graphics card, which is selling like hot cakes but is hard to find, turns out to be in such a shady place..."

Attached with the photos of the mining farm computer room sent by the mine owners: in a remote and simple computer room in a mountainous area, rows and rows of densely packed Zhiyun OPA1000 graphics cards are working hard!

Soon, a netizen made a joke about Zhiyun graphics card: Not good at gaming, but great at mining!

This situation was also noticed by Xu Shenxue. After all, he was also a young man who went online for entertainment every day and read the news.

Not good at gaming, but great at mining. This comment left Xu Shenxue speechless, although he also agreed with the netizens' opinions. However, if people from the GPU business group under Zhiyun Semiconductor saw this, they would probably feel complicated.

This means that after a year or two of hard work, what they finally came up with is still a GPU computing power card...

If that's the case, why are they working so hard? It would be much easier to just castrate the AI ​​chip and push it directly to the market.

Alas, it seems that this is not possible, because the current AI chips are still in a highly confidential period and cannot be announced to the public at all. Currently, only the group’s intelligent research institute and some special domestic institutions can access and use AI chips.

But no matter what, launching consumer-level computing power cards for the market is not the original intention of Zhiyun Semiconductor GPU Business Group. They plan to make gaming cards, targeting a large number of game players at home and abroad.

The current situation is not what they want!
Of course, it’s a good thing that the graphics cards can be sold... although they don’t make much money.

The independent graphics card market these days is actually very small. Even with many additional businesses, the total market size is only a few billion.

Nvidia's annual revenue is only over 4 billion, and ATI's is even less, only over 1 billion US dollars.

Nvidia's market value has been only a few billion dollars for a long time, and it has even dropped slightly in the past two years.

And recently, affected by Zhiyun's entry into the consumer graphics card field, the stock price of Nvidia next door has also seen a small decline. It is almost unable to maintain its market value of 8 billion US dollars... The main reason is that Zhiyun's name is quite scary. After Zhiyun entered the independent graphics card market, it affected the sentiment of Nvidia investors.

Just when Zhiyun Semiconductor's GPU business group was still worrying about how to open up the gaming graphics card market.

However, Xu Shenxue has already summoned the investment department of Zhiyun Group to discuss the feasibility of acquiring NVIDIA.

The people in the investment department didn't understand why their big boss would be interested in a "small business" that mainly produced gaming graphics cards... But since the boss had spoken, they still made a very detailed report and gave their own suggestions.

"NVIDIA's current core business is high-performance graphics processors, mainly supplying GPU processors to various graphics card manufacturers, and also mainly providing some development software!"

"Although the company has a relatively large advantage in the vertical segment of the discrete graphics card market, with a market share of more than 70%, it is limited by the small capacity of the discrete graphics card market and the overall growth of the PC market has shown a downward trend. Its overall revenue is relatively small and its growth potential is limited."

"The company's revenue this year is estimated to be around $4.2 billion, with a net profit of around $600 million, and a current market value of $8.3 billion."

"According to the data we collected, their revenue in the fourth quarter may have only increased by two or three percentage points year-on-year, but their operating expenses increased by at least 10% year-on-year, resulting in a year-on-year decline in net profit of about 15 percentage points. This is a very dangerous signal."

"Based on the overall environment of the current independent graphics card market, it will be difficult for them to get rid of this situation in the short term, even within two to three years, and it will be difficult for them to significantly increase their revenue and profits."

"At the same time, it does not have any new businesses or products that can turn the situation around. The company is highly focused on the field of high-performance GPUs, which certainly brings them great advantages in the vertical track of discrete graphics cards, but also brings them huge limitations. It is almost impossible for its core business to transform!"

Senior analysts in the investment department are obviously not optimistic about Nvidia, and various analytical data also support their views.

Xu Shenxue's analysis and predictions about them were also anticipated. Nvidia is in such a weak position now.

But who could have imagined that the future development of AI would depend on a tiny graphics card. Nvidia, which makes graphics cards, took advantage of the AI ​​boom and saw its revenue soaring rapidly while its profits soared even more!

With a revenue of 23 billion US dollars in years, the net profit can reach billion, and the net profit ratio is close to %.

It's quite exaggerated. Its ability to make money is even better than that of high-tech giants like Zhiyun.

Zhiyun always says that it has a strong ability to make money, but the gross profit of its iconic S series mobile phones is only over 50%, and the net profit is about 30%.

The net profit ratio of Zhiyun's entire smart terminal business is only about 25%.

Don't underestimate the 25% net profit ratio. It is already a very good ratio in the high-tech manufacturing industry. Especially for a company as large as Zhiyun, it is rare to maintain such a high net profit ratio.

At the same time, this net profit ratio and Zhiyun’s huge sales volume are also the core factors supporting Zhiyun’s huge market value.

But in the era after the explosion of AI, NVIDIA's net profit ratio reached 50%...which is even more awesome than Zhiyun's current one.

This ability to make money can be said to be quite explosive!
And everyone is optimistic.

In the early 2000s, as long as you recognized AI, you basically had to recognize Nvidia's market value, because Nvidia's AI chips were almost equivalent to AI at this stage.

A number of technology giants are trying various ways to purchase Nvidia's AI chips, and some are even begging Nvidia to accept their money and then be kind enough to sell them AI chips. This was said by the boss of Oracle himself.

This shows Nvidia’s strong dominance and importance in the future AI era... How important is it?
Not to mention AI chips, even the consumer-grade gaming graphics card 4090 has been banned from sale!
However, these current investment analysts are certainly unable to predict such future events.

Although the investment analysts in Zhiyun Group's investment department are all elites in the industry, they also have their limitations. They can only make judgments based on various existing data and industry development trends, and such judgments are extremely limited and even short-sighted.

Well, if they could accurately predict the development of the future situation, they would not have come to Zhiyun to be investment analysts, but would have started their own business to develop AI chips.

Therefore, Xu Shenxue was not surprised that they were not good-looking, nor did he blame them for having poor taste.

He just said: "I have almost understood the details, so there is no need to say more. Now I have a question, can we acquire it?"

At this time, several senior analysts in the investment department looked at each other for a while, then lowered their heads to discuss for a while, and one of the senior analysts said: "If the price is right and a certain premium is offered, there is still a possibility of acquisition, but cash acquisition is somewhat difficult, mainly because there may be some non-business obstacles, and at the same time, its management and major shareholders may not be willing to give up the company for cash."

"It should be smoother if it is through mergers and acquisitions. After all, our company's stock is still very popular. Their shareholders should be more willing to exchange Nvidia's stock for our Zhiyun's stock!"

"There will probably be some obstacles from non-commercial factors, but as long as they are not made too loud and threatening, then it won't be a big problem."

People in later generations may think that NVIDIA is particularly important, but in 13, NVIDIA was actually not that important, even inconspicuous.

"After all, in the current PC GPU market, the mainstream is still Intel and AMD's integrated graphics cards, and Nvidia's independent graphics cards have a relatively low market share and influence."

"In the current PC GPU market, NVIDIA's integrated graphics have an absolute market position, accounting for more than 70% of the market, especially in the fields of notebooks and office computers, where Intel's integrated graphics have an absolute advantage!"

"AMD occupies part of the integrated graphics market, and also occupies part of the discrete graphics market."

"The remaining ones are Nvidia's independent graphics cards, which have a relatively small market share and a small impact. Therefore, if nothing unexpected happens, they should not be subject to obvious interference or resistance!"

As he spoke, the investment analyst also made a witty remark: "After all, independent graphics cards are just tools for some video game enthusiasts to play games. They don't receive much attention overall, and their impact on the industry is also small. No big shots pay attention to them... except for you, Mr. Xu!"

There is one point that the investment analyst did not mention, which is that these data only apply to the GPU field on the PC side. In fact, there is also a mobile GPU field that seems to be ignored.

In this regard, Zhiyun is the king, followed by Qualcomm, and then Apple... basically all of them are mobile phone SOC manufacturers.

It’s just that people generally separate the PC market from the mobile market, so no one thinks Zhiyun is a big player in the GPU market...even though it really is. Every year, the shipment volume of GPU cores built into SOCs in the S series and W series reaches hundreds of millions, far exceeding Intel’s integrated graphics.

So don’t think that Zhiyun has no strength in the GPU field just because it didn’t make any independent graphics cards before. In fact, in the GPU field, Zhiyun’s scale far exceeds that of Intel and Nvidia, and its technology is no worse than theirs.

It’s just that mobile GPUs are not getting much attention.

Xu Shenxue said: "Since this is a possibility, then we should take corresponding actions. We can give an appropriate premium on the price, but I have one requirement, which is to acquire 100%. We can even adopt a merger and acquisition model and exchange some stocks of our Zhiyun or its subsidiaries."

"I want to have 100% control of this company, and then..."

Then many operations are not suitable to be told to these investment analysts.

However, we can refer to the post-acquisition processing of many overseas companies. After the acquisition, they obtained numerous patents and technical information, and then established R&D centers in China, forming a R&D structure with the R&D center in China as the main and overseas R&D centers as the auxiliary.

After the acquisition, although the overseas subsidiary is still headquartered overseas, it will actually gradually become a business subsidiary responsible for sales, after-sales and technical support.

If the acquisition of Nvidia is successful, the same operation will be adopted... Nvidia will continue to exist and even continue to launch new gaming graphics cards every year.

It’s just that the R&D center will gradually be transferred to China, and Nvidia’s subsidiary in the United States will become a purely business subsidiary.

At the same time, Xu Shengxue will personally cut off its AI chip project in its overseas R&D center... It cannot be said to be cut off, but it will be merged into the GPU business group of Zhiyun Semiconductor.

Zhiyun Group does not need a second AI chip project, and the world does not need a second one... At least before the AI ​​explosion, Xu Shenxue does not want to see any second AI chip project.

The AI ​​chip can be kept secret for as long as possible. When it can no longer be hidden, the company can use the name of Nvidia's subsidiary to launch outdated versions of AI chips to overseas markets to compete for the overseas market, while suppressing other overseas GPU chip companies from making AI chips.

From the beginning to the end, Xu Shenxue's goal was AI chips...

Whether they are developing consumer-grade graphics cards themselves or trying to acquire Nvidia, the purpose has never been the gaming graphics cards themselves, but AI chips!

As for the gaming graphics card business... Xu Shenxue never took it seriously. The entire market capacity is only a niche market at the level of several billion dollars, and Xu Shenxue was too lazy to make any plans.

After all, you can't really expect Xu Shenxue to go out and do 4090 for a group of gamers...

After explaining the task to the people in the investment department, the people in Zhiyun's investment department also began to take action. In order to ensure confidentiality and to minimize interference from outside the business, Zhiyun did not contact them directly, but contacted several major shareholders and management through intermediaries.

The reason for the acquisition or merger is not AI chips, nor is it because they are interested in independent gaming graphics cards, but to integrate the PC business.

Now Zhiyun’s PC business, CPU and various chipsets are being developed.

There is also integrated graphics, although it is not very good.

Not to mention the memory and flash memory, which are world-class.

Among all the core components in the entire PC business, the only missing one is a high-performance independent graphics card. This is why the outside world believes that Zhiyun Semiconductor is making efforts in the field of independent graphics cards.

Unfortunately, although the OPA1000 released by Zhiyun Semiconductor seemed to be selling well, those were just a few sales supported by mining bosses, and were not considered by the outside world to be a real success.

The OPA1000 graphics card has very poor gaming support performance and is obviously unable to achieve Zhiyun Group’s ambitions for the PC business.

After NVIDIA learned of Zhiyun Group's acquisition intention, it conducted a lot of internal analysis and basically agreed with this statement... If it weren't for the PC business, it would be impossible for them to be interested in a company that makes independent graphics cards.

The question is, to sell or not to sell?
If you want to sell, at what price should you sell it? Should you go through the acquisition model and take the cash and leave, or go through the merger and acquisition model and take Zhiyun’s stocks and follow Zhiyun from then on?

They themselves are also very conflicted about this!
(End of this chapter)