Rebirth 08: Rise from copycat phones

Chapter 157 Domestic CPU and AI Chips
Currently, the communication baseband project that Xu Shenxue has seen at Zhiyun Semiconductor... there is one, but that is a baseband chip based on VMware Telecom's original CDMA2000 communication baseband. This thing is actually okay. Although it is not very successful in commercialization, there will be no major problems after simple improvements.

Then we also made WCDMA and TD-SCDMA baseband chips for testing... The technology used in these two types of communication baseband chips was not very advanced, and the communication basebands produced were not very good.

The mobile phone team has tested these two communication basebands before, and the signals were so bad that they were basically unusable.

The signal is poor and unstable!
But there is no other way but to keep working on it and try to come up with a good communication baseband chip as soon as possible.

At the same time, we are also striving to design a baseband chip for all networks... After all, we have collected all the patents for the three major 3G standards, and it would be unreasonable not to develop a full network support.

After taking a quick look at the PPTs, Xu Shenxue said, "We still have to keep working on the baseband chip project. If we don't come up with it soon, 3G will be outdated!"

"4G technology is already being promoted outside now!"

Xu Shenxue's words made many people in the baseband project feel embarrassed... Damn, we have been working on 3G baseband for more than a year, but there is still no result... If it is delayed for a few more years, just like Xu Shenxue said, 3G will be outdated and we will have to use 4G by then.

Compared with its backwardness in 3G, Zhiyun Technology's layout in the 4G field is earlier and much smoother.

In the first half of this year, we signed a licensing agreement for the fourth-generation communication protocol with Japanese operator NTT Docomo. This is FDD-LTE... This thing is one of the main branches of the famous 4G technology and is also an international mainstream technology.

Prior to this, Zhiyun Technology also signed a cooperative development agreement with China Mobile on the fourth-generation communication technology, namely TD-LTE technology, which is also the main branch of 4G technology and is mainly used domestically.

So although 4G technology has not yet been popularized and there are only a few test networks at present, Zhiyun Technology has actually obtained patent authorization for two mainstream 4G technologies.

Compared with the difficulties in obtaining patents in the 3G era, it is much simpler and easier to obtain 4G technology.

Zhiyun Technology is also the first company in the world to collect patented technologies for communication basebands of many current network standards... from 2G GSM to the three major 3G network standards, and then the two 4G network standards...

As for Zhiyun Technology, based on the patents it owns, it is planning to develop a full-network communication baseband... and integrate communication basebands for various network standards such as 2G, 3G, and 4G.

We no longer have to develop a separate model for each network standard as we do now, which is very troublesome.

But it's not easy to get a full network communication baseband...

-

Be it mobile phone SOC or communication baseband, there are achievements, but they are not too many. For the time being, even until next year, they cannot fully support the development of Zhiyun Technology.

We still need to keep working hard!
Of course, Xu Shenxue is still patient about this. After all, he already has the patents, so there is no need to worry too much about being stuck. He can just get back to it and continue working on it.

After looking at these two major projects, Xu Shenxue also looked at some other chip projects of Zhiyun Semiconductor.

Zhiyun Semiconductor's chip business, although mobile phone SOC and communication baseband are its core businesses...well, these two will actually be integrated into the same project in the future.

At the same time, there are also many other chip projects, such as some mid- and low-end chips used in many products in the ecological chain...for example, there are also memory chips and flash memory chip projects.

However, the progress of these two methods is not smooth. It will take some time to prepare, and then acquire some related technology companies from outside... If we rely on ourselves, patents will be a big problem.

In the chip field, especially the more mature CPU/GPU as well as memory, flash memory, and communication baseband, patents are very strictly controlled.

If newcomers want to enter the industry purely through their own research and development without acquisition, it is almost impossible... At the very least, you have to get some patent licensing.

The best way is to simply acquire a company that has relevant patents and technologies and start from that basis.

In the above fields, Zhiyun Technology basically follows this path...

Otherwise, the patent issue cannot be avoided...and what Zhiyun Technology is doing is not something that can be seen in the light, and it has to be sold globally, so the patent issue must be complied with.

Failure to comply with patents will be very troublesome. Not only will you have to pay huge sums of money in lawsuits, but you will also be denied entry into some foreign markets.

Therefore, the patent issue must be resolved...and there is no good way to resolve it except to spend money!

Of course, there are many chip projects that do not have so many messy restrictions. For example, the patents of low-end chips used in many ecological chains are easy to solve. If all else fails, you can develop them yourself, which is not a big problem.

For example, the small-scale chips used in smart watches and the chips used in tablet computers actually all use the ARM architecture. Zhiyun Technology itself has permanent authorization for the relevant patents and can tinker with them as it pleases.

The ARM architecture is an open architecture. It is very easy to use and it is based on the Android core, so the ecological chain can also be easily solved.

Xu Shenxue was thinking about whether he could make an ARM-based PC chip and use it as a notebook chip, paired with a PC system modified from Android...

In the original time and space, Fruit used AMR chips and carried its own MAC OS system... It also adopted a large amount of IOS ecological content in the ecological chain... It achieved a seamless transition and ultimately achieved huge success.

If Fruit can follow this path, Zhiyun can do the same!
It is impossible to bypass X86 these days. If you want to create a simple new architecture and a new ecosystem, it will be equally difficult... The ecological chain is a huge problem.

However, if you use the ARM architecture, combined with the Android core modified system, and use the existing Android ecosystem, the overall difficulty will drop rapidly... Of course, to do so, the software ecosystem must also provide strong support and adaptation solutions.

Otherwise, all you’ll end up with is a large Android tablet!

Xu Shenxue has some ambitions in this regard, but he still needs to consider carefully how to do it... After all, there are no small problems when it comes to ecological issues.

It is not a big problem to develop a desktop-level CPU, and it can also be done by converting the YUN OS system to a PC system.

However, how to make a large group of software vendors under the YUN OS system willingly and even proactively improve and adapt the software to adapt to the PC system is a huge problem!
So we can’t rush this, we have to take it slow… For example, we have to develop a lot of developer software first to reduce the difficulty and cost of adaptation for software developers. ————

In addition to this, Zhiyun Semiconductor has also accepted research and development commissions from some companies... Some companies want to obtain purely domestic CPUs for absolute security reasons, and then install domestic Linux secondary development systems... As for ecological issues, they are not civilian consumer products, so there is no need to consider any ecological issues. A few customized software are enough.

Before this, domestic chip design companies were basically not very good, until Zhiyun Semiconductor emerged... Then some people took a look and thought, wow... Zhiyun Semiconductor's technology seems to be quite awesome!
Then they came to me and asked if they could make a CPU for computers...the kind with its own instruction set.

This... this has stunned the people at Zhiyun Semiconductor. Everyone on earth knows that we are working on the ARM architecture... We have never worked on an independent instruction set and have never had this idea before!
But since they came to us and expressed clear purchasing needs, Xu Shenxue gave the instructions and just went for it... It was a good opportunity to try out the most basic chip design technology, the autonomous instruction set.

But it will take a long time to develop this thing... The design of a completely new independent instruction set alone cannot be perfected in one or two years.

In addition, Zhiyun Semiconductor is also developing computing cards...well, or graphics cards.

This computing card is not intended for use by the ecological chain, but for the Intelligent Research Institute... Since its establishment, the Intelligent Research Institute has been working on the so-called intelligent AI all day long, and then they proposed the need for a more specialized, high-performance chip for big data processing.

This kind of chip doesn't exist on the market at all... Then they tried to use a graphics card to do the calculations, and found that it worked very well.

But ordinary civilian graphics cards are still somewhat insufficient. They need large professional-level graphics cards with higher computing power, or computing cards.

Therefore, Zhiyun Semiconductor took on this task and prepared to develop a professional big data computing card with extremely high computing power for the Intelligent Research Institute... and this thing has a new name in later generations, called AI chip.

But from a hardware perspective, the so-called AI chip is actually an oversized graphics card...

Nowadays, no one would have thought that a graphics card that was mainly used to play games or, at most, perform various video image processing... would eventually become a core productivity tool of a new era... A lot of new technologies would rely on this thing.

The AI ​​of the new era refers not only to AI drawing, AI novel writing, AI question-and-answer and other insignificant things...these are just the most superficial applications.

In fact, there are also intelligent driving unmanned vehicles and intelligent industrial manufacturing in the civilian field.

There are also a large number of unmanned equipment and AI command centers in the defense field, such as unmanned fighter jets, unmanned ships or tanks, mechanical dogs, war command AI, etc.

All of these rely on large-scale model calculations to train the algorithms, and this training process is extremely dependent on AI chips.

Why did they ban your AI computing card, and even the civilian graphics card 4090? It’s not because they are afraid that you can play games, or draw pictures with AI, or even write novels with AI.

People are afraid that you will use this thing to train various industrial AIs and national defense AIs.

Why are we afraid? Because we really do it... and we do it very well... When other countries are using AI to do AI conversation, painting, and text editing, we are using AI chips to do smart manufacturing and national defense AI...

While other companies' robot dogs are still performing somersaults to boost stock prices, our robot dogs are already carrying rocket launchers for live-fire testing.

Then the other side was dumbfounded... and quickly banned the AI ​​chip, and even did not allow you to use chip foundry.

The typical case of unplugging the network cable when you can’t handle the game anymore!

And Xu Shen was concentrating on developing a semiconductor manufacturing backup plan... just in case the other party pulled out the network cable!
----

There are so many chip R&D projects, which has also made Zhiyun Semiconductor grow bigger and bigger, with thousands of R&D engineers...

After leaving Zhiyun Semiconductor, Xu Shenxue called Bai Qiwen from Wecoo Electronics and asked him to develop a new series of mobile phones to be equipped with the newly developed S100 chip from Zhiyun Semiconductor... and said that Zhiyun Semiconductor could supply the products at cost price. He was not seeking to make money, but only to promote large-scale testing to gain experience and feedback!
When Bai Qiwen heard about this good thing of purchasing at cost price, he naturally agreed without hesitation... As for the possibility that domestic chips would lead to the decline of the brand, he didn't have to worry about it at all.

The two brands under Weiku Electronics have never relied on any big brand spare parts for advertising... Weiku Electronics' advertising strategy is actually very simple, just telling consumers: My big brother is Zhiyun...

Then in reality they rely on low-priced domestic spare parts, and then add some imported backward chips, and fiddle with the so-called cost-effectiveness, make the appearance a little prettier, and then sell it cheaply!
Relying on this set of methods, Weiku Electronics' monthly shipments hit a new high after this summer. In June, relying on the popularity of low-end large-screen mobile phones, the shipments exceeded four million units.

Although the phones sold are all mid- and low-end, the monthly sales volume of more than four million units is still a very large number.

For a company like Weiku Electronics, it would be a great idea to install some domestic SOCs on it and sell it with the best cost-performance ratio.

After finalizing this matter, Xu Shenxue also shifted his attention from the chip field and re-invested in the semiconductor manufacturing field!
Although chip design is very important, it is just a castle on the beach... Even if the chip you design is the most amazing, as long as the other party does not provide foundry services and you do not have the corresponding independent technology, the unregulated chip foundry will be doomed.

In the 21st century, especially after the 1990s, the entire modern technology is largely based on semiconductor technology.

Not to mention the various smart terminals, let’s talk about the computing cards that are very important in the future... This thing has very high requirements on the manufacturing process, and the top-level computing cards often also require the top-level manufacturing process technology.

All of these are based on semiconductor manufacturing technology.

Therefore, compared with the field of chip design, Xu Shenxue actually pays more attention to the field of semiconductor manufacturing...

It is not just about chip foundries, but about the entire semiconductor industry chain, including the upstream semiconductor manufacturing equipment and materials fields.

To be more precise, Xu Shenxue pays great attention to anything that may be a bottleneck... So no matter whether these projects make money or not, Xu Shenxue will do it in advance and put a lot of effort into it.

He used part of the US$5 billion he cashed out from his shares to acquire various semiconductor manufacturing equipment and materials-related industries, and the rest was used to establish various new startups and then invest in research and development of related basic industries.

By the way, his R&D investment in the field of semiconductor manufacturing equipment and materials alone exceeds the total R&D investment in these basic fields in the country over the past decade!

Xu Shenxue is serious about basic research and development!

(End of this chapter)