Rebirth 08: Rise from copycat phones

Chapter 488 Commercialization Plan of Virtual Equipment
Xu Shenxue has imagined many times the scenes of himself using virtual devices. He also knows that current science and technology cannot support the kind of immersive virtual devices in novels that can provide five senses... Current brain-computer technology is still in a very primitive stage.

In addition, Xu Shenxue did not expect that he would be able to produce a virtual device with the same technical level as Ready Player One at the moment.

Apart from anything else, the physical sensation alone cannot achieve it.

At present, the virtual devices that Zhiyun Group’s virtual device project team can produce can only provide visual and sound immersion, while enabling free movement and adding a certain amount of force.

However, the freedom of movement and force provided by the omnidirectional treadmill, somatosensory suit, and force bracket are still relatively primitive. Especially in terms of force, they can only simulate a simple degree of force and feedback it to the user through physical means.

Just looking at the performance parameters and features of these devices, Xu Shenxue actually didn't have much hope for this second-generation prototype at first.

However, after actually experiencing it and being in this wonderful 3D starry sky scene, Xu Shenxue was still shocked!

All of this looks so real!
As the starry sky scene changed rapidly, Xu Shenxue soon found himself on a beach. Although he could not feel the softness of the sand, he could see the sparkling sea. The sea water repeatedly rushed up to the edge of the beach and then retreated, creating waves of white waves.

Xu Shenxue then tried to squat down and looked carefully at the sand and sea water at his feet!

You can see the sand texture and sea water on the surface.

He found that his feet were still several centimeters sunk into the sand... Even though he couldn't feel it, it still felt novel to see it now.

He tried to reach out his hand to scoop up some sand... and he actually scooped up some sand. At this time, he even felt some touch... This was the result of the large number of micro motors deployed in the special gloves in the somatosensory suit, which provided the user with a certain amount of pressure for simulation. Although this was a very shallow and even fake pressure touch, it was still a major breakthrough.

However, he looked at the sand carefully and touched it with his fingers, and found that the sand was still grains.

This made him feel very magical... Could it be that the engineers really made a pile of sand?
This is impossible. Even if it can be done, the computing power cannot support it.

Then he tried to scoop up other sand, and after careful observation, he found that the sand looked the same... and its movement posture was almost fixed, and would not change much due to his random movements.

Well, this is just a simple copy and paste, but even so it is not easy.

Then he touched the sea water again. Although he did not feel the coldness of the water, he saw the movement of the sea water caused by his fingers. More importantly, he also found that his hands were stained with water...

Just this image of sea water that looks so real; the sand that can be picked up and then falls through the gaps between fingers; and the image of fingers touched by water, are enough to reflect Zhiyun Group's strong technical accumulation in the field of graphic display technology over the years.

Even before the experiment, the engineers said that when they were designing the graphics, they used long-distance mapping and close-range physical simulation.

In other words, only the sand and water around Xu Shenxue look real, and the sand a little further away is just a texture...

However, this still made Xu Shenxue feel very surprised!

At least the scenes he had seen so far had exceeded his expectations!
The graphics are super high definition, and the physical collision simulation is more realistic, which is a huge leap compared to the first-generation prototype he experienced last year.

------

Then Xu Shenxue stood up and began to try walking. He felt that walking was not much different from usual. It can be seen that the omnidirectional treadmill performs quite well.

Extending from here are all kinds of the most advanced technologies available today.

The movement of the omnidirectional treadmill itself is actually nothing. The most core technology is how to capture the movement posture of the person, and then mobilize the omnidirectional treadmill to perform corresponding directional changes, so that the user can walk easily and naturally.

This involves top-level motion capture technology, and motion capture technology involves lidar and other sensor equipment!

This is also the reason why Xu Shenxue put on a body sensing suit in addition to the omnidirectional treadmill's own laser radar and high-definition camera. The main purpose of this body sensing suit is to collect his movement posture data.

Finally, this requires very powerful computing power support. The more sensors there are and the more precise they are, the more data will be generated.

The more data there is, the more computing power is required to process it.

When the technicians introduced it before, they said that in order to process such a huge amount of sensor data, they specially used multiple ZY38 chips to form a sensor data processing array, so that they could process such a huge amount of data almost instantly and then feedback the results to the artificial intelligence center of the entire system.

The central computing center, built on the underlying technology of artificial intelligence, will assign tasks to the motion center system based on these motion data. The motion center system will then mobilize the universal treadmill and semi-fixed frame to ensure natural and smooth use by users.

At the same time, sports data will be fed back to the game center and synchronized with the game content.

The most core underlying technology here is the collection and instantaneous processing of motion data, each of which represents the highest level of contemporary human technology.

----

After Xu Shenxue walked a few steps and felt good about it, he tried jogging, running fast, and made some changes of direction and emergency stops, and he felt there was no problem with it.

Xu Shenxue spent a lot of time watching the pictures and experiencing various movements. The overall feeling was that it was a great improvement over the first generation of virtual equipment, with great improvements in both pictures and movements.

In Xu Shenxue's opinion, the current level of technology is already fully in line with his imagination of virtual games, or the virtual world.

Of course, there are still problems... This thing uses an omnidirectional treadmill. If the character wants to exercise, he has to move by himself, which is not good for people with average physical strength.

But this is not a big problem and can be changed through game content, such as assigning characters with means of transportation to avoid having to walk.

Things like flying boats, flying carpets, and flying swords can easily solve this problem.

Afterwards, Xu Shenxue officially started the game, entered the game world and began various shooting operations.

The biggest feeling Xu Shenxue had was that the immersive experience of the game was very good, and it even gave people the illusion that this was not a virtual world, but a real world.

Xu Shenxue focused on the game graphics, which were done very well, so well that it was difficult to tell the real from the fake.

The movement system is also very well done, and can perform various movements very naturally.

These experiences made Xu Shenxue, who had never played games much, play for more than 20 minutes without realizing it.

After the experience, Xu Shenxue was still a little dazed when he took off his helmet... Not long ago, he was fighting alien monsters in a city ruins, but now he was in a bright white laboratory.

It feels like virtual and reality are intertwined!
It's amazing!

After the staff helped him take off the suit, Xu Shenxue asked directly without waiting for the staff to speak: "Tell me, how long will it take for this thing to be mass-produced, and how long will it take to bring down the price?"

"To be honest, the second-generation prototype has made a qualitative leap compared to the first-generation prototype. I think with the current technical level, it can be put into commercial use!"

At this time, Qu Xuewen, the executive president of Zhiyun Group, who is responsible for a series of business lines of virtual equipment, said: "Thank you, Mr. Xu, for your affirmation. At present, if our second-generation model is to be mass-produced, it is actually not a problem. Although we have adopted a lot of new technologies, it is not a problem to achieve mass production!"

"The core components are advanced sensors and computing chips. The AI6000 graphics card array provides the overall computing power of the entire system. We use four AI6000 graphics cards to form the basic central computing array, but we have also tested the use of APO5000 graphics cards. If this type of computing power card is used, the overall performance will not drop too much!"

"After all, the APO5000 graphics card is just a scaled-down version of the AI6000 graphics card. The two have the same architecture, but the memory capacity and communication bandwidth are different. Our system is not used for artificial intelligence training, but only runs a small dedicated artificial intelligence built into the terminal, so it is not a big problem!"

"In addition, we also used the ZY38 data computing chip, which was originally intended to be installed in smart robots and other medium and large terminal products for external sales. We can use it directly without making any changes."

"Both of the above chips are produced using an equivalent 7-nanometer process. Although they are expensive, mass production is not a problem. Moreover, our project team is only a small user of these two chips, so their costs will not be affected!"

“Sensors and other spare parts can also be mass-produced!”

"If necessary, we can definitely achieve mass production!"

"But it is more difficult to reduce the cost. When we calculated the internal hardware cost before, we only calculated the internal purchase price of these chips. For example, for the AI6000 graphics card, we only calculated the internal purchase price of one hundred thousand. If we use the APO5000 graphics card later, the cost will be no less than or ."

It is worth noting that the scaled-down version of the AI6000 graphics card, APO5000, has not yet been officially launched to the public, but if nothing unexpected happens, the launch price will be very high, and may exceed RMB per card in the initial stage of launch.

After all, the previous generation APO4600 graphics cards, which use a -nanometer process, are now in short supply, and the price on Zhiyun's official website remains at yuan per card.

The more advanced APO5000 will naturally be more expensive!

The selling price of Zhiyun Group's computing power card has little to do with the actual cost, and is mainly based on market pricing... because the actual manufacturing cost of this thing is very low. The pure hardware cost of APO4600 is only more than 5000 yuan, and APO is more expensive, but only more than yuan.

However, Zhiyun Group cannot sell according to cost, but rather according to market supply and demand!
This is also the fundamental reason why the gross profit margin of Zhiyun Group's APO series computing power cards can reach more than 90%: supply exceeds demand!

Even the AI ​​series graphics cards purchased and used internally by Zhiyun Group are very expensive... After all, although many project teams belong to the group, each project team also has its own KPI, and it is impossible to let other project teams use their own results for nothing.

Each subsidiary and project team also has its own pressure.

Therefore, even for internal procurement, the cost cannot be really calculated. A certain profit must be given to the sister subsidiaries and the project team.

Qu Xuewen continued: "In addition to the APO5000 graphics card, the price of the ZY38 chip is also relatively expensive."

The ZY series chips are dedicated sensor data processing chips provided by Zhiyun Semiconductor specifically for medium and large artificial intelligence terminals.

"The new ZY38 chip has more than billion transistors and is relatively expensive. Although we have intelligent robot projects, Hailan car projects, low-altitude drone projects, and other artificial intelligence terminal projects to share the cost, the internal supply price is still per chip. In this system, we use a total of four chips!"

Xu Shenxue frowned slightly when he heard this. Four APO5000 graphics cards alone, at a price of 38 yuan each, would cost yuan. Even if calculated based on the internal cost price, it would cost more than yuan... Two ZY chips would cost yuan in total.

Just for two types of computing power cards, the internal purchase price costs more than 400,000! If they are put on the market for sale, it is impossible to use this internal purchase price to calculate, but to calculate according to the external wholesale price.

Zhiyun Semiconductor's graphics cards, which are priced at RMB 200,000 each, are being sold well and are in short supply... You use a price of RMB 100,000 to calculate the cost and then sell them to the market. Zhiyun Semiconductor must be so pissed!
This means that the cost of these two chips will skyrocket to more than 900,000!
Qu Xuewen continued: "In addition to these two core chips, our entire system also has CPU chips, storage chips, memory chips, and an unreleased X5 gaming graphics card that specifically provides image processing for helmet displays!"

"And then there are a large number of various auxiliary chips in the system!"

"For the entire system, we use about three thousand chips of different types!"

"In addition to chips, there are also various hardware devices. The most expensive ones are various sensors, including lidar, body sensing suits, gloves, and various sensors built into shoe covers!"

"The hardware cost of the chip and various sensors is estimated to be around 1.8 million, and the rest is the cost of the machine, display, treadmill and other equipment!"

"In the end, the hardware cost of various spare parts is about two million, but this does not include our R&D costs, nor the assembly costs of the production line!"

"Our R&D budget is huge. If we calculate based on an annual production of 10,000 units, and include R&D costs, various production costs, and administrative costs, the selling price must be at least 5 million per set to break even. If we want to make a profit, we need to further increase production and sales, because the higher the production and sales volume, the lower our R&D and manufacturing costs will be!"

"According to our previous calculations, if we can achieve an annual production and sales volume of 30,000 units, we will be able to lower the price to 3 million per set!"

"If the production and sales volume can reach 50,000 sets, we can even lower the price to 2 million per set!"

"Of course, the above prices do not include profits, we can only say that we are not losing money!"

"If we take into account profits, especially the profits from sales to maintain subsequent research and development, the overall price will have to increase by more than 30%!"

Xu Shenxue listened briefly and looked at the cost statistics table brought out by the staff next to him.

Five million per set, this price is ridiculous. It is impossible for it to enter the consumer market, and there is not even much of a corporate market.

If the price can reach two million and then further reduce to around one million, Xu Shenxue believes that there will be a certain market, and it will not be a personal consumer market, but an enterprise market!
For example, some companies purchase these devices, and then set up virtual game operation venues, provide rental services, virtual game experience services, etc.... probably, it is a bit similar to the Internet cafe model.

For example, some training institutions purchase these devices and then use them as various simulation devices, such as for learning to fly various aircraft and various vehicles.

There is another more important potential customer, that is the military customer!
This is the most realistic and most likely customer to purchase such expensive equipment at the moment!

Military customers have also experienced it before. Last year, Xu Shenxue sent them a batch of first-generation models for experience, and their feedback on the experience was very good. It is very useful for various training.

Now the second-generation model is out, with even better performance, and will be even better for training.

But no matter what the customer is, the price must be lowered first... a price of several million is bound to be how much it will be sold for.

After thinking for a while, Xu Shenxue said, "According to the cost calculations you listed, the core costs are all kinds of chips, especially all kinds of chips with advanced technology, such as seven-nanometer chips and ten-nanometer chips!"

Qu Xuewen said: "That's right, our entire system is actually built on artificial intelligence technology, and the underlying hardware of artificial intelligence technology is the computing chip."

"So, Mr. Xu, you can see that these computing chips account for almost half of the cost of all our components!"

Xu Shenxue said: "Then can we use slightly outdated computing chips? What about the 4600-nanometer APO34 graphics card and ZY chip? After the next year, the production capacity of these two chips will be very sufficient, and the internal supply price will drop significantly!"

Qu Xuewen said: "Of course this is possible. When we conduct research and development, we use these two chips for research and development supporting tests, but the overall performance will be greatly reduced, perhaps only about 50%."

Xu Shenxue asked: "How about comparing it to the first generation model?"

Qu Xuewen said: "Of course it is better than the first generation model. Our second generation model has not only upgraded the computing chip, but also a large number of other hardware as well as the underlying AI algorithms and model frameworks!"

Xu Shenxue said: "In this case, if we use these two mature chips, the cost can be controlled. In addition, we can continue to use the current X5 graphics array and maintain the image quality, then the overall problem will not be a big problem. The slight lag in sports posture performance will not be felt by ordinary people. At least for me personally, the difference between the second generation and the first generation of sports posture is not that big!"

After Qu Xuewen heard Xu Shenxue chopping a chip, he then heard Xu Shenxue talking about some other chips and sensors, especially the expensive ones, one by one, asking if they could be replaced with mature products!

What else can Qu Xuewen say? He can only say that both are possible... After all, in fact, both are possible!
But after all the hacking, Qu Xuewen looked at the new parts and components, and then said: "Mr. Xu, this is not much different from the first-generation prototype. Except for the helmet display and graphics card using new technology, the rest are basically the same configuration as the first-generation model!"

Xu Shenxue said: "Just tell me if this can be used!"

Qu Xuewen said: "Of course it can be used!"

After thinking for a while, he continued, "And with our technological advances in various algorithms, even if we use the hardware configuration of the first-generation model, we can still produce products with performance that exceeds the first-generation model!"

Xu Shenxue said: "That's what I mean. Since the bulk of the cost is various advanced chips, we have to take into account the fast cycle characteristics of the chips!"

"Two or three years ago, 12-nanometer chips were still top-level technology and expensive, but now they are very cheap!"

"The same goes for the 300-nanometer chips. They are too expensive now. We can't afford them. So we can wait until next year. Next year, our 5000-nanometer production capacity will increase to 38 pieces per month, and we will use the HEUV-B lithography machine to produce them. The cost will be greatly reduced. Then you can use the PO graphics card and ZY chip!"

“We must look at technological progress and cost reduction from a development perspective!”

Before, Xu Shenxue was fooled by the high hardware cost of this second-generation model.

But he soon realized that the price of chips would drop, and the speed of price reduction was very fast!

The prices of the new generation of chip processes are very inflated. This is because the supply and demand are in excess, which leads to inflated prices. At the same time, semiconductor factories will take advantage of the inflated prices caused by the supply and demand to recover costs as quickly as possible.

A semiconductor factory with advanced technology requires an investment of tens of billions of dollars. Semiconductor companies will naturally want to recover their costs as much as possible during the period of advanced technology.

If you delay for a few more years, you may lose everything!

Because the pace of technological advancement in semiconductors is very fast and new processes are iterated quickly...this also leads to very high depreciation costs for various equipment in semiconductor factories.

The equipment depreciation costs for other types of factories are often calculated over five or even ten years.

However, the equipment depreciation calculation period for advanced semiconductor factories is often only about three years.

The semiconductor industry is quite special. Advanced technologies are in short supply, while mature technologies are in plentiful supply and no one wants them.

This is also the reason why semiconductor manufacturers are working hard on research and development, trying their best to maintain themselves in the first echelon and strive to be the first!
This is because the first tier will take up more than 90 percent of the profits of the entire industry, and the top-ranked company will take up at least half of the profits.

The particularity of the semiconductor industry also leads to another result, that is, the price of chips with mature processes will fall rapidly.

The current APO4600 graphics card is still priced at RMB per piece, but by next year or the year after, the price will drop significantly, and it is estimated that it can be purchased for around RMB .

The price of APO2.5, which uses 4500D packaging, old-generation architecture, and worse memory and communication bandwidth, has now dropped to or yuan. It is estimated that it will be available for or yuan by next year.

However, in contrast, the APO5000, which is expected to be launched at the end of the year, will have an extremely exaggerated price, directly reaching around yuan.

However, virtual devices are not for training artificial intelligence, and they do not require the latest graphics cards... Those who are chasing the latest graphics cards are major companies, and they do so in order to train artificial intelligence and deploy and operate artificial intelligence.

Therefore, they are willing to spend huge sums of money and place orders of hundreds of thousands of graphics cards that cost $200,000 each. They only ask for quick delivery!
However, there is no need to rush Zhiyun's own smart terminals. It is possible to make do with the APO4600 graphics card first, and then use a large number of mature chips and equipment to make a virtual device with better performance than the first generation, and control the cost within one million. This is the best solution in Xu Shenxue's opinion.

"This virtual device is divided into different levels. The flagship model uses the current second-generation advanced configuration, which has all the good stuff and is priced higher."

"Then we will use mature chips and spare parts to make a low-end version, and keep the price under one million!"

"This is the only way to open up the market. The equipment costs several million per set. Ten sets of equipment means an investment of tens of millions. Even a virtual Internet cafe cannot be opened!"

“Price is the key. By the way, other unnecessary costs should also be saved to further control costs!”

"And don't spend too much on R&D costs. The R&D costs have already been invested, so we can calculate them slowly later!"

"Control the hardware cost and plan a gross profit of about 30%. Start the business first, then we can gradually expand sales, spread the cost, reduce the hardware cost, and enter the profit stage!"

Qu Xuewen was naturally a smart man. When he heard what Xu Shenxue said, he said, "In this case, we can also imitate the operating strategies of some intelligent robot projects. They adopt the method of not making a profit in the early stage in exchange for greater sales, thereby spreading the cost and then making a profit!"

Xu Shenxue said: "However, the robot project is doing very well. Now the hardware business has also started to make a profit, and the profit is very good. Why is their hardware business profitable? Because the annual sales volume of robots has reached more than one million units, which greatly dilutes various costs!"

"Sufficient sales are the basis of everything!"

“Without sales, there is no revenue, and without revenue, it is meaningless to talk about gross profit margin!”

Qu Xuewen said: "Later, I will lead the team to make a plan for you, Mr. Xu, to review."

Xu Shenxue said: "Well, plan well, find out more about the marketing department and cooperate more. They are the professionals in selling things!"

A few days later, Qu Xuewen found Xu Shenxue to report on the newly released commercialization report of the virtual device.

(End of this chapter)