Rebirth 08: Rise from copycat phones

Chapter 370: The Shocking Robot Conference
Xu Shenxue returned to the Zhiyun Group headquarters and convened a small-scale meeting with several group executives to discuss the 5G technology standards.

"At present, we still have some shortcomings in the patent layout in the 5G field. We need to further cooperate with more manufacturers to implement a model of advancing and retreating together, so that we can promote our technical standards!"

"But even so, it will be very difficult. Europe has already decided to support Siemens' technical standards, the United States is headed by Qualcomm's standards, and here we are headed by ourselves."

"It is now a three-way battle, and each party has its own advantages and disadvantages. However, considering the facts, it is impossible for us and them to get rid of each other and do their own thing. After all, adopting a unified 5G standard is our common demand and the main reason why we sit down with them for negotiations!"

"For long codes in channel coding, we mainly promote the domestic Polar code solution, but this standard has little support overseas. This is mainly because our company and several other domestic manufacturers have accumulated a large number of patent layouts in this regard, and overseas manufacturers naturally do not agree to use it."

"The same is true in reverse. For the LDPC coding scheme promoted by manufacturers such as Quadstar and Qualcomm, they have more patents, while we have relatively few patents in this field. If we adopt the LDPC coding scheme, we may have to pay certain additional patent fees."

"In terms of short codes, it is a three-way battle. The technical differences in each aspect are not big. At the same time, we have gathered enough manufacturers to support us, so we have a big advantage in short codes!"

“So if nothing unexpected happens, the long code will use the LDPC standard, and the short code will use the Polar standard. This is also the result that many manufacturers are happy to see!”

"The problem now is that the number of patents held by each party is different, and patent exchange is needed, but the current patent exchange negotiations are not going smoothly."

Xu Shenxue said: "Let's continue the discussion. I have only one request. No matter what standard we use, we will not pay even a dime in patent fees!"

The so-called dispute over standards, to put it simply and bluntly, is actually a dispute over patent fees!
A number of Chinese companies, led by Zhiyun Group, are unwilling to pay patent fees to Western companies led by Samsung and Qualcomm... Of course, Western manufacturers are also unwilling to pay patent fees to Chinese companies led by Zhiyun Group.

Therefore, in this competition for 5G standards, it is impossible for any one company to dominate and for one standard to become the global standard... because the other party will simply not pay attention.

Wanting to collect money just by relying on a so-called technical standard is not even possible in dreams.

If we insist on doing it by force, then everyone will do their own thing, that is, two 5G standards will be developed, but this is actually unlikely.

Because both parties have made patent arrangements in these two 5G standards.

For example, Qualcomm and Quadstar are promoting the LDPC standard. They have many patents in this technical standard, but Zhiyun Group only has fewer patents in the field of LDPC solutions, not that it has no patents.

Even among the patents of the LDPC standard owned by Zhiyun Group, many are core patents. If you want to abandon Zhiyun and develop the PDLC standard yourself, Zhiyun will block you...

The reverse is also true!

This means that neither side can actually abandon the other and develop a separate set of 5G standards, and patent issues will be very troublesome.

If we develop two 5G standards and cross-exchange patents with each other... this is purely unnecessary. It would be better to exchange patents in advance and then develop the same 5G standard.

The 5G technology standards are extremely complex, involving more than patents. They have formed a situation where you are in me and I am in you... In fact, no one can abandon anyone else.

This is also why major manufacturers sit down to negotiate and try to reach a unified standard... because they can't do it alone.

Regarding the complexity of the 5G standard, Xu Shenxue’s instruction was that no matter what standard is used, no patent fee will be paid.

It would be even better if you could charge some patent fees instead of paying any... Of course, the other party is not stupid either.

Therefore, the negotiations between the two parties are still ongoing, and the outside world has also seen that although many manufacturers have tried to initiate 5G standard meetings and then conduct negotiations, they have all been postponed by Zhiyun Group.

Without even negotiating a good patent exchange or distributing benefits, Qualcomm and Snapdragon 888 want to call a group of overseas manufacturers to a meeting and vote to push the LDPC solution without authorization?

If they dare to mess around, Zhiyun will use the core patent of the LDPC solution to block them!
Therefore, Zhiyun postponed the meeting, and all parties accepted it reluctantly, waiting for Zhiyun to continue negotiations with Qualcomm, Sixin, Siemens and other companies.

When the negotiations are successful, the meeting can be held to vote... Before that, there is no point in throwing a large number of Chinese manufacturers led by Zhiyun Group to vote to push the LDPC solution.

Especially Zhiyun Group. Even if they don't use patents to block you, as a core member of the 5G Technical Standards Committee, the so-called technical standards without the participation of Zhiyun Group are not global standards...

Of the 1.5 billion smartphones shipped worldwide each year, half of them are equipped with Zhiyun Group's SOCs, and these SOCs are equipped with Zhiyun Group's self-developed B series communication baseband.

In other words, half of the mobile phones sold worldwide each year use the communication baseband produced by Zhiyun Group's semiconductor subsidiary.

If Zhiyun doesn’t agree, what’s the point of playing with 5G!

Of course, the same can be said for several overseas manufacturers such as Qualcomm and Quadcore... Without their approval, 5G cannot be implemented.

This is a dispute over interests among seven or eight giant companies in the communications industry. It is complicated and simple at the same time. The core is nothing more than patent fees.

Xu Shenxue's attitude towards the 5G standard is very clear: let's talk about it slowly, and no patent fees will be paid anyway...if the negotiations fail, just drag it out.

Anyway, 4G is quite useful now. Zhiyun makes the same profit by selling 4G phones and 5G phones...it doesn't affect anything.

For users, the key factor affecting Internet speed is not 4G or 5G, but the data throughput of the operator's backbone network... That is a matter of the fiber optic backbone network, and it has a certain relationship with wireless communication technology standards, but not much.

Traffic charges also depend on the operator's investment scale and operating costs, and have little to do with wireless technology standards... In fact, new technology standards will lead to new large-scale investments, and traffic costs will be even more expensive.

In addition, the construction of 4G in China has only just started, and the huge investment has not yet recovered the cost. It is too early to talk about 5G now.

As for the Internet of Things, 4G is actually quite useful, and even 3G is not bad... When Hailan Automobile first started working on car networking, it had a built-in 3G communication baseband that seems outdated now, but it was just as useful for uploading/downloading data.

Even 3G/4G signals are better... 5G's advantages are low latency and high speed, but its communication distance is a shortcoming. It is easy to lose the signal over a long distance...

This is also why in the 5G era, 5G base stations are everywhere and investment has increased significantly, because the 5G signal communication distance is short and the coverage area is small, so more base stations are needed to complete the coverage.

Therefore, there is no need to rush when it comes to 5G. Wait a few years, until the investment in 4G in China is almost recovered, and then we can slowly develop 5G.

------

Compared with promoting 5G, Xu Shenxue pays more attention to the various applications of artificial intelligence technology, which is the killer weapon that can truly change people's lives and bring about industrial upgrading.

In early December, Xu Shenxue once again attended an artificial intelligence technology conference and gave a brief speech at the conference, talking extensively about the field of artificial intelligence and advocating the various advantages of artificial intelligence technology.

As specific carriers of artificial intelligence technology, such as HaiLan Automobile's autonomous driving technology, robotics technology, and voice assistant technology on smart terminal devices, were all highlighted by Xu Shenxue.

At the same time, Xu Shenxue also mentioned some underlying technologies of artificial intelligence technology, such as neural network data big models, and revealed to the public for the first time some underlying technical principles of Zhiyun Group's artificial intelligence technology. He also announced that several data models will be open source to support overseas and domestic manufacturers to enter the field of artificial intelligence.

Open-sourcing several data models is the result of careful consideration by Zhiyun Group.

The main purpose is to promote the development of the entire AI field... and then sell chips!

After all, one of Zhiyun Group’s core businesses is the semiconductor business!
This year, Zhiyun Group has launched a series of AI chips based on 12-nanometer process, and will launch several AI chips based on -nanometer technology in January.

At the beginning of this year, Zhiyun launched the APO4000 graphics card based on the first-generation 4500-nanometer process, and in January next year it will launch the APO graphics card based on the -nanometer process.

At the same time, there is also the NVIDIA-branded X graphics card in the gaming field, and the X3 graphics card based on the nm process will be launched in the near future. In addition to being a high-end gaming graphics card, this graphics card can actually be used as a low-end AI training graphics card. Many small and medium-sized manufacturers and small research institutions use this graphics card for AI training.

Then the NVIDIA brand under Zhiyun Group also has a PX airborne computing power platform that can be used in fields such as autonomous driving or robotics. The PX computing power platform based on the 14-nanometer process was just launched in August this year.

In order to boost chip sales in these AI fields, Zhiyun Group chose to open source several algorithms... Of course, they are all unimportant open source models, and are mainly concentrated in primary fields such as human-computer interaction and image recognition.

It is a field that is easier for people to access and for Internet companies to engage in.

As for the real core technologies of artificial intelligence, especially some core algorithms, they are the technologies that Zhiyun Group has kept in its hands and it is impossible for them to be open source.

Zhiyun Group plans to promote the development of the entire AI industry by opening up several AI algorithms, thereby boosting the sales of AI chips. This is mainly because Zhiyun Group's sales in the field of AI chips have continued to increase in the past few years, and it has established a huge technological advantage. At this time, it is natural to hope that larger manufacturers will participate and sell more chips.

"If nothing unexpected happens, this year our revenue from AI training chips, including APO series graphics cards and X series graphics cards, will exceed 8 billion US dollars. This field will become a new important revenue growth point for our group, and the profit in this field is extremely high. The gross profit margin is expected to reach more than 5 billion US dollars, which will significantly increase the group's overall net profit!"

"In the field of airborne computing chips, our PX computing platform is gradually expanding the market and has developed very well in the field of autonomous driving. Many domestic and foreign customers have adopted our PX platform to develop autonomous driving technology."

"At the same time, we will release a new airborne computing chip under the APO brand in January next year, focusing on the low-power APL series, which will be used for small and medium-sized robots or other intelligent devices."

"Through dual-brand and dual-series operations, we can cover a wider market and meet the needs of different customers!"

Fu Zhengyang, senior vice president of Zhiyun Group and general manager of Zhiyun Semiconductor Co., Ltd., described the prospects of the AI ​​chip field to Xu Shenxue: "This time, by open-sourcing some algorithms, we will continue to promote the development of artificial intelligence technology at home and abroad. Our entire AI chip business segment is expected to exceed US$13 billion next year!"

“It has become the fastest growing and most profitable business in the group’s semiconductor business.”

Xu Shenxue looked at the report and said, "This is good. The overall decision is good. Keep going like this. Remember, we are not only providing artificial intelligence technology, but also providing the underlying technology of artificial intelligence!"

“But if we want to really push forward the development of AI technology, we have to take drastic measures!”

After thinking about it for a while, Xu Shenxue thought of the generative AI that was being developed in the group… Although conventional artificial intelligence is very important and is also a core technology, people’s feelings are still not strong enough.

Xu Shenxue plans to provide a powerful AI technology, namely generative AI, which is AI that generates various texts/pictures, to a large number of companies, especially ordinary people, next year.

In fact, these things have already existed within Zhiyun Group. Strictly speaking, the series of image-based technologies deployed by Zhiyun Group on smart terminals, such as automatic beautification, automatic photo editing, image recognition, and tagging, all belong to this set of technologies.

However, Zhiyun Group did not pay much attention to this before. After all, Zhiyun Group is not engaged in the Internet, but mainly in hardware, so it is not very interested in generative AI, and therefore has not invested too much resources in it.

However, since we want to promote the development of the entire AI track and stimulate chip sales, we can still do more in this regard.

Thinking of this, Xu Shenxue went to the Institute of Intelligent Research that afternoon to take a look at the development of their generative AI.

After a simple trial experience, I feel that it is better than the previous version of generative AI.

"This thing is great. Now that it's out there, don't keep it to yourself. Find a suitable time to release it next year so that ordinary people can experience AI technology and see if we can boost the AI ​​chip market!"

What is the most expensive cost in the AI ​​era?
It’s not the development of the model itself. The salaries of a few hundred or a few thousand engineers don’t actually cost that much... The real cost lies in computing power.

The computing power required to train a large model is enormous, and a super server consisting of tens of thousands or even hundreds of thousands of APO graphics cards is needed for training.

After the large model is trained, if it is to be used for public services and to provide services to a large number of users, it will require huge operational computing power... This also requires a super server composed of a large number of APO graphics cards.

Computing power requires hardware costs. APO graphics cards are not cheap, costing tens of thousands each. In addition to hardware costs, there are also electricity costs. Maintaining the operation of the server, cooling, etc. all cost money.

So it is actually very expensive to engage in AI these days.

This is also the reason why Zhiyun Group has developed a generative AI with relatively good performance a long time ago, but has never put it into commercial use... For such a large generative AI model, the computing power required for subsequent operation is very expensive, let alone training.

Even if we charge users, it will be difficult to maintain!
But now, with the continuous advancement of technology, although this kind of generative AI is still very expensive, it has some commercial value.

Besides, if we don’t do it ourselves, we can let other high-tech companies do it!

Zhiyun itself can just sell an APO graphics card...

Therefore, Xu Shenxue decided to launch this generative AI into the market next year as a prototype, and after causing a sensation in the market, launch another wave of generative AI open source... Tsk tsk, it is hard to say whether those companies engaged in generative AI will make money by then, but Zhiyun Group's APO graphics cards will definitely sell well!

How can you make more money by digging for gold than by selling shovels?

----

While Xu Shenxue was thinking about making generative AI open source and then selling AI chips, Zhiyun Group's marketing of robots was still continuing and continued to increase in popularity over time.

It was not until mid-December, on the eve of Zhiyun Group’s winter press conference, that various discussions on the field of robotics reached their peak.

On the day of the press conference, people who had been repeatedly bombarded with various news began to pay attention to Zhiyun Group's winter press conference.

They want to see whether Zhiyun Group has really come up with a groundbreaking black-tech robot, or whether it is just bragging.

Obviously, the marketing department of Zhiyun Group has the confidence to brag so much and spark a global discussion.

On the evening of the press conference, Xu Shenxue and a group of group executives arrived at the press conference, and a container vehicle also came with them!

So, this scene appeared in the cameras of a large number of media on the scene: when Xu Shenxue got out of his vehicle, he came to the front of the container car.

After the staff opened the door of the container and lowered the slide connecting to the ground, they saw several wheeled robots walking down from the carriage.

Then eight bipedal robots came down.

Eight wheeled robots and eight bipedal robots got out of the carriage and lined up behind Xu Shenxue.

The scene, and everyone present, was shocked...

At this time, Xu Shenxue smiled and said to a large group of media cameras: "Welcome to the new era of robots!"

In just a few minutes, some self-media reported the news first, with a very exaggerated title: Xu Shenxue and his robot army!
When netizens who were watching the live broadcast online saw this scene, they were even more shocked. Is this robot serious?
It looks pretty cool.

"Is this a fake, or is it real?"

"It looks like a showpiece, just like the robot released by Zhiyun before. It looks very cool at first glance, but it can only pour water. It is basically useless!"

“It doesn’t really matter what the robot looks like, what matters is what it can do.”

At this time, at the press conference, Xu Shenxue continued: "Let's go to the press conference together!"

As he spoke, Xu Shenxue walked along the aisle into the venue, and the eight wheeled robots and eight bipedal robots behind him also lined up to follow.

Outside the passage, hundreds of media reporters went crazy when they saw this scene. They kept taking pictures, and some anxious ones even shouted while trying to conduct interviews.

"Mr. Xu, are these robots operating autonomously or remotely controlled?"

"Mr. Xu, what are these robots used for?"

However, at this time, Xu Shenxue obviously would not answer any questions from reporters. This scene was originally part of the press conference. Let the robot walk out directly from the container car and then follow Xu Shenxue into the venue, giving people a visual shock and facilitating subsequent publicity.

Xu Shenxue brought a group of robots to the conference venue. After he stepped onto the stage, he began to say, "Welcome everyone to our Zhiyun Group's winter conference!"

"You should all be very interested in the robots behind me. Yes, as you can see, this is the universal robot YunBotot developed by Zhiyun!"

“A great product that represents the highest technological achievement of contemporary mankind!”

(End of this chapter)