Rebirth 08: Rise from copycat phones

Chapter 641 The Sixth Generation Superconducting Quantum Computer
On October 17, at a regular press conference of Zhiyun Group, the company's spokesperson publicly acknowledged for the first time that Zhiyun Group possesses a superconducting quantum computer that can be applied on a large scale, and claimed that the company has already applied the superconducting quantum computer in multiple artificial intelligence computing centers.

It also acknowledged for the first time the high efficiency of superconducting quantum computers in the fields of artificial intelligence training and operation.

Although the spokesperson for Zhiyun Group did not provide any detailed technical information at this press conference, especially crucial performance parameters such as the number of qubits, the news was still quite sensational.

As soon as the press conference ended, major online media outlets around the world were the first to report on it, and a large number of independent media bloggers quickly followed suit, with various information about Zhiyun Group announcing its superconducting quantum computer filling the entire cyberspace.

In no time, netizens around the world who pay even a little attention to the technology industry saw the news on various social media apps, especially short video platforms, news sites, and Weibo.

This shocked many people!

Quantum computers, which have been touted in science fiction for years and repeatedly researched by the industry, have finally entered the practical application stage?

Many people who are knowledgeable about industry dynamics or who are focused on quantum computer research and development have also expressed their views.

An American television news website invited an expert in the field of quantum computing to ask questions.

The expert stated on the television program that it is incredible that Zhiyun Group can produce a practical quantum computer at this stage, because there are many companies or research institutions researching quantum computers, but they have not been able to solve the problem of the number of controllable qubits even in the laboratory... let alone commercial applications.

However, he also stated that if breakthroughs are achieved in multiple underlying technologies, it is entirely possible for Zhiyun Group to develop a superconducting quantum computer.

It is also speculated that Zhiyun Group may be using big data models to predict and intervene in qubits, thereby achieving large-scale controllability of qubits.

It was also pointed out that this is currently the technical route being promoted by many companies researching quantum computers.

Given Zhuyun Group's significant technological advantages in the field of artificial intelligence, it is entirely possible for them to do so.

And his guess was correct... Zhiyun Group's quantum computer did indeed rely on its powerful artificial intelligence technology to create the LX algorithm architecture, train a big data model specifically for predicting and analyzing qubits, and thus achieve large-scale intervention and analysis of qubits.

Only in this way can we achieve tens of thousands, or even hundreds of thousands, of controllable qubits, allowing quantum computers to enter a truly practical stage!
The core technology of Zhiyun Group's superconducting quantum computer is still based on artificial intelligence technology... Of course, conversely, the emergence of superconducting quantum computer technology has further promoted the development of Zhiyun Group in the field of artificial intelligence technology.

These two are interdependent and have a spiraling upward relationship.

Sigrún has taught at the Iceland University of the Arts as a part-time lecturer since and was Dean of the Department of Fine Art from -. In – she held a research position at Reykjavík Art Museum focusing on the role of women in Icelandic art. She studied fine art at the Icelandic College of Arts and Crafts and at Pratt Institute, New York, and holds BA and MA degrees in art history and philosophy from the University of Iceland. Sigrún lives and works in Iceland.

There are also numerous reports and discussions about superconducting quantum computers on many domestic media websites. On Zhihu, the number one topic is: How exactly does Zhiyun's superconducting quantum computer work, and what kind of impact will it have?
There were a lot of answers to the topic, with several thousand appearing in a short period of time. Of course, the vast majority of them were meaningless jokes and memes, and many were emotional posts. There were very few answers that actually contained any information.

One of the answers analyzed the situation from a professional perspective, attempting to explore the technological roadmap of Zhiyun Group's superconducting quantum computer... Its analysis results were similar to those of overseas professionals, both identifying the technological breakthrough as: using big data models to predict and intervene in quantum bits!

"Zhiyun Group must have developed some kind of big data model that can accurately predict and intervene in qubits; otherwise, it would be impossible to significantly increase the number of controllable qubits in a quantum computer!"

The blogger also made some speculations, suggesting that Zhiyun Group has made a significant technological breakthrough in quantum chip manufacturing.

"Currently, the core computing units within quantum chips are achieved through stacking, and this stacking precision needs to reach the hundred-nanometer level. According to publicly available information, the number of stacked units in quantum chips made using this technical approach is currently unsatisfactory!"

"To further increase the number of core computing units inside a quantum chip, the stacking precision must be improved. Only in this way can more computing units be stacked on the quantum chip, which is something that current technology cannot achieve!"

"Zhiyun Group must have found a completely new process route, thereby enabling the manufacture of quantum chips with more computing units!"

This blogger is speculating on a new technological approach... In reality, there are some differences. Zhiyun Group adopts a completely new self-developed 3D stacking technology approach, for which it has rebuilt an entire production line and developed a variety of core manufacturing equipment, the most important of which is the quantum alignment machine.

Its fourth-generation quantum alignment machine can achieve alignment offset at the level of 20 nanometers, which is an astonishing level of precision. Currently, apart from Fairy Mountain Holdings, which possesses such an ultra-high precision quantum alignment machine, there is no other company on Earth with similar technology.

Let alone the fourth-generation quantum alignment machine, even the alignment offset of the previous third-generation quantum alignment machine at forty nanometers is a level of precision that no other company's equipment can achieve.

The fourth-generation quantum alignment machine is just one of the core pieces of equipment in the quantum chip production line. There are more than ten other types of core equipment, and more than one hundred high-precision equipment and more than two hundred auxiliary equipment.

The entire quantum chip manufacturing plant is actually more complex and technologically advanced than a three-nanometer chip factory!

The manufacturing of advanced quantum chips is inherently more difficult than that of semiconductor chips manufactured using a three-nanometer process.

Zhuyun Microelectronics, a subsidiary of Zhuyun Group, can now mass-produce chips using the 3-nanometer process, with a yield rate of several tens of percent.

However, Zhiyun Group is still unable to achieve large-scale, low-cost mass production of advanced quantum chips... even the fifth-generation quantum chips that have already been mass-produced... have very low production capacity, with an annual output of less than two thousand chips!
The main product is the superconducting quantum chip from Zhiyun Group. In order to meet the special requirements of the LX architecture, a brand-new 3D superconducting component stacking quantum chip architecture was specially developed.

Its technical approach and processing method are fundamentally different from those of quantum chips processed by other manufacturers using traditional semiconductor technology!

These two are not the same thing!
To manufacture a 3D quantum chip, Zhiyun Group needs to precisely stack hundreds of thousands or even more components on the chip, and the alignment offset must be controlled within tens of nanometers each time. If the stacking deviation of a single component is too large, it will cause the entire component's working area to fail. To address this, Zhiyun Group's quantum chip design department has specially designed a functional area shielding technology to shield the failed functional areas that occur during the manufacturing process.

However, if this happens too many times, causing too many functional areas to fail, the entire quantum chip will still be rendered unusable... Even the strategy of shielding the failed functional areas will not work, as these functional areas are highly interconnected. Shielding a few functional areas is possible, and it can still be used with difficulty... but the number of controllable qubits will decrease.

However, if there are too many shielding elements, then there's no way to fix it; it will directly render the entire quantum chip unusable.

Of course, if you are extremely lucky and there are very few or even no faulty functional areas in the chip manufacturing process, then the number of controllable qubits will exceed the design expectations. For example, in the fifth generation of quantum chips, the top-of-the-line quantum chip can actually reach 12,000 qubits... but the inferior products only have about 8,000 qubits.

Even though they are both fifth-generation quantum chips, there are performance differences!
However, regardless of the method, the number of components that need to be stacked and processed in the 3D stacked quantum chip developed by Zhiyun Group is too large, and the precision requirements are extremely high, which ultimately leads to a very low yield.

The extremely low yield rate means low production capacity. At the same time, the R&D costs of quantum chips need to be factored in, as do the expensive investment costs of building quantum chip factories. Ultimately, this results in the current quantum chip costing extremely high amounts, sometimes tens of millions of dollars per chip.

In order to produce quantum chips, Zhiyun Group, together with Fairy Mountain Holdings, specially developed a complete set of production equipment, which is currently unmatched by any other factory.

------

While the outside world was hotly debating the news about Zhiyun Group's superconducting quantum computer, the superconducting quantum computer project team within Zhiyun Group was also making improvements to the YCL-51 superconducting quantum computer according to instructions from higher-ups... or rather, weakening it!

The YCL-51 superconducting quantum computer is intended for external sale... The group's senior management is unwilling to sell it directly. The performance of this superconducting quantum computer is too powerful, and selling it directly would easily break their computing power advantage... At the same time, the MMK34 material cannot be disclosed to the public and must be replaced.

Therefore, the MMK34 material used in the YCL-51 superconducting quantum computer needs to be completely replaced with TK120 superconducting material... This will lead to greater cooling requirements, so the heat dissipation system needs to be redesigned.

Of course, it's just a heat dissipation system, and the overall technical difficulty is not great... Especially for a top company like Zhiyun Group, they don't need to redesign the core cooling system; they can simply use the cooling system from their previous fourth-generation quantum computer.

Then, the fifth-generation quantum chip was weakened by simply and crudely cutting off a large number of components... Fewer components mean simpler processing, saves more time, and can even significantly improve yield, thereby reducing manufacturing costs and increasing profits.

The YCL-51 superconducting quantum computer costs 300 million yuan to purchase internally alone. If it is to be sold externally and still maintain a profit, it will cost 600 million or even 1 billion yuan.

To be honest, not many companies can afford this price... So the price needs to be controlled while maintaining profits. Therefore, we first cut MMK34 superconducting material and replaced it with cheaper TK120 superconducting material. At the same time, we also need to cut the quantum chip price!

A fifth-generation quantum chip costing tens of millions of yuan is too expensive and the cost is too high. It would be best to make a quantum chip with a cost of less than ten million yuan.

This export version of the fifth-generation quantum chip, with its weakened design, was designed to have between four and six thousand controllable qubits. The main reason for this is that during the actual manufacturing process, alignment accuracy deviations resulted in inconsistent shielding areas, thus limiting the number of controllable qubits to between four and six thousand.

If the number of qubits is less than four thousand, it means that a large number of functional areas of the chip are disabled or blocked, making it unusable.

With 6,000 qubits, this means that the design limit has been reached, and only a very small deviation occurred during the entire manufacturing process.

Going forward, Zhiyun Group will classify each chip into different levels based on the condition of its component functional areas, resulting in quantum chips with four thousand, five thousand, or six thousand controllable qubits, and will also set different prices for them!
After reducing performance requirements, the improved yield will eventually lead to a significant decrease in cost. According to the project team's estimate, the cost of the weakened fifth-generation export version of the quantum chip is expected to be controlled at five million yuan per chip.

Apart from the major modifications to the superconducting materials (with the accompanying cooling system) and the quantum chip, the modifications to other subsystems were minor. They were basically simple modifications that saved some costs before being put into use.

This design of a directly weakened quantum computer is relatively simple to implement because it does not involve technological breakthroughs... especially for a top company like Zhiyun Group.

Not long after, Xu Shenxue saw the first prototype.

Its size is similar to that of the YCL-51 quantum computer, both being large cabinets and not small in size... Of course, Xu Shenxue doesn't care about that; he cares about performance and cost.

The head of the quantum computer project team reported to Xu Shenxue: "The quantum computer we modified, internally codenamed YCL-52, has 5,000 controllable qubits. There are also versions with 4,000 and 6,000 qubits!"

"Using TK120 superconducting material as the core material!"

"Its overall performance is far from comparable to our own YCL-51 superconducting quantum computer. If used for artificial intelligence training, its efficiency is only about 40%!"

"However, the cost has indeed been reduced significantly. After taking all costs into account, we estimate that the cost of each unit is around 50 million yuan. In addition, the cost of the 4,000-qubit and 6,000-qubit models is the same as that of the 5,000-qubit models, with no significant difference!"

"As for the selling price, that depends on the marketing department's pricing strategy and the external market environment!"

Xu Shenxue nodded after listening: "That's right, although the performance is a bit lacking, it's enough to meet market demand!"

He wasn't really that concerned about the YCL-52 superconducting quantum computer... it was just a monkey version for export, used to make money and also for exchanging benefits with those major conglomerates.

His real focus was on the next-generation sixth-generation superconducting quantum computer: YCL-61.

This is the core project of the quantum computer project team, as well as the other subsystem development project teams!
Its core technical indicator is only one: the number of controllable qubits has been increased to 50,000!

Compared to the fifth-generation YCL-51, it is five times better!
Achieving this was no easy feat. Zhuyun Group designed and built a new generation of quantum chip manufacturing equipment and production lines, enabling them to produce sixth-generation quantum chips with a significantly increased number of components.

At the same time, the core algorithm also needs to be significantly upgraded and improved in order to meet the prediction and intervention requirements of more than 50,000 qubits.

Neither of these two core technologies is simple.

Even though the quantum computer project at Zhiyun Group is staffed with top geniuses and has various levels of research opportunities, it is still not easy to accomplish.

Since the YCL-51 superconducting quantum computer completed its technological development, they have started working on the sixth-generation superconducting quantum computer, but it still took more than three years to complete.

Now, Xu Shenxue has finally seen the mass-production prototype of the sixth-generation superconducting quantum computer!
Its appearance and size are not much different from the fifth generation. The appearance is still a very simple big iron box. Because it was for Xu Shenxue to see, one side of the cabinet door was open, and Xu Shenxue could see the many complex devices inside!
Although the core component of a superconducting quantum computer is the quantum chip, which is actually very small, it still requires a large number of external components and subsystems to support its operation... and it cannot be miniaturized yet!

Therefore, the overall size of a quantum computer is relatively large!

Engineers at Zhiyun Group are currently exploring the miniaturization of quantum computers, but it remains quite challenging at present…

Current superconducting quantum computers are actually very similar to early electronic computers. Both are still in the early stages of development, and they are very large and consume a lot of electricity.

However, given time, especially with the emergence of higher-performance room-temperature and room-pressure superconductors, future superconducting quantum computers may be as small as personal microcomputers... and with further development, they might even be able to fit into laptops or even mobile phones.

That's a long way off in the future... Right now, the superconducting quantum computers developed by Zhiyun Group are behemoths!

The head of the quantum computer project also stated, "We have completed the mass production prototype of the sixth-generation quantum computer and are currently preparing for mass production. It will be used in the second YANC computing center!"

"The second YANC computing center will fully utilize our YCL-61 superconducting quantum computer, which will further enhance the overall computing power!"

The first YANC computing center in Xiyang City is using the YCL-51 superconducting quantum computer to control technological risks, and because the mass production time and speed of the YCL-61 superconducting quantum computer cannot keep up. It is expected to be put into operation next spring.

However, the second YANC computing center, which will also be built in Xiyang City, will use the YCL-61 superconducting quantum computer and is expected to be put into operation in the summer or autumn.

When the third YANC computing center is built, it will be further upgraded, such as using APO7000 graphics cards with equivalent three-nanometer process and fourth-generation high-bandwidth memory. This computing center is expected to be operational next winter.

Barring unforeseen circumstances, Zhuyun Group will acquire and put into operation three new next-generation YANC computing centers next year, adding 1,000 EFLPOS of AI computing power, bringing Zhuyun Group's total AI computing power to 1,600 EFLPOS.

Four more computing centers are expected to be put into use the following year. With the deployment of the new generation of neuromorphic computer chips, along with the YCL-61 superconducting quantum computer and the APO7000 graphics card, the computing power of the newly built YANC computing center will be even greater, and it is expected to add 2,000 EFLPOS of artificial intelligence computing power the following year.

Going forward, Zhiyun Group will maintain an annual increase of more than 2,000 EFLPOS in AI computing power until the domestic power generation capacity can no longer support it... and this figure is approximately 10,000 EFLPOS.

At that time, if we want to continue to significantly increase computing power, we need to make the existing power supply provide as much computing power as possible through continuous technological innovation... However, due to the limited speed of progress in basic technologies, such an improvement will be difficult.

Another option is for China to significantly accelerate the construction and supply of traditional power facilities, but this is also difficult... Traditional power facilities, such as hydropower and nuclear power, have long construction cycles and high investment costs. It's impossible for Zhiyun Group to simply provide as much power as it wants...

At the same time, it is impossible to shut down the power supply of other companies just to meet the power needs of Zhiyun Group. Other industries are also very important!

The last option is controlled nuclear fusion, which can provide a large amount of electricity.

Xu Shenxue's passion for controlled nuclear fusion and building a lunar base is actually largely driven by the desire to provide his company with more electricity!
He didn't realize it before, but in the last two years he has increasingly felt the limitations of the power supply!

Without a sufficient power supply, many things cannot be done.

Of course, this is a happy problem for Xu Shenxue... He only has so much electricity because he works on artificial intelligence in China, which allows him to run such a large-scale computing center.

If it were any other country, it would be impossible for them to provide him with hundreds of billions or even trillions of kilowatt-hours of electricity... It's not that they're unwilling, but that they can't.

Other overseas companies, such as Google, Microsoft, and a large number of other companies involved in artificial intelligence, are now also facing power shortages... and the solutions are much more complicated than those for Xu Shenxue.

Once the competition in artificial intelligence reaches a certain scale... if the technologies of both sides are roughly the same, then it is actually a competition of electricity!

In this respect, China has an absolute advantage!

Generating a lot of electricity is awesome!
(End of this chapter)