Rebirth 08: Rise from copycat phones

Chapter 596 First Flight of Project X3
At Gulf Technology, Xu Shenxue first inspected the core manufacturing equipment for the sixth-generation superconducting quantum chip: the fourth-generation quantum alignment machine.

Then I looked at several other core devices needed to manufacture quantum chips!

The entire process of manufacturing superconducting quantum chips places top-level requirements on the industrial system that are no less demanding than those for manufacturing advanced semiconductor electronic chips. Both represent the pinnacle of current human-made micro-manufacturing technologies.

In order to build the entire manufacturing chain and develop various top-notch production equipment and consumables, Fairy Mountain Holdings has invested heavily in the past few years and deployed a large number of R&D personnel to develop the entire manufacturing chain.

And this advanced quantum chip industry chain is the first of its kind in the world, there is no other like it!

Of course, this refers to top-tier technology, with quantum chips capable of manipulating tens of thousands of qubits... If it's just some quantum chips developed by other companies that only have a few dozen or at most a few hundred qubits, that's not difficult. They can be made using existing equipment, and some quantum chips can even be made using existing semiconductor equipment.

However, quantum chips with only a few dozen or a few hundred qubits have extremely limited uses and can only play a role in a very few specific fields. They cannot be used as the core computing power for artificial intelligence, as is the case with companies like Zhiyun Group.

100,000 qubits versus hundreds of qubits – this is the huge technological gap between Zhiyun Group and others in the field of quantum computing.

This technological gap is also the core strength that enables Zhiyun Group to continue to lead the world in the field of artificial intelligence.

When it comes to artificial intelligence, besides algorithms, it's all about competing on computing power.

In terms of computing power, superconducting quantum computers have a huge advantage... not to mention that Zhiyun Group has already developed a neuromorphic computing system. If this technological advantage is also added, the future computing power advantage will be even greater.

The dual advantages of algorithms and computing power are the core factors that enable Zhiyun Group to gain such a significant technological edge in the field of artificial intelligence and apply AI technology to many areas.

Superconducting quantum computers are crucial to the group's computing power and its leading position in artificial intelligence technology, so Xu Shenxue attaches great importance to them.

After observing the research and development and production of various core equipment for quantum chips at Haiwan Technology, Xu Shenxue took the opportunity to hold a seminar on quantum chip equipment and consumables in the conference room of Haiwan Technology.

In addition to top-tier equipment system integrators like Gulf Technology, the conference also included more than twenty quantum chip equipment and consumables core subsystem or material R&D or suppliers under Fairy Mountain Holdings.

More than 20 senior core technology management personnel from various subsidiaries discussed the research and development and production of a series of equipment and consumables required for the manufacturing of sixth-generation quantum chips at the seminar.

When faced with difficulties in research and development or production by various subsidiaries, Xu Shenxue would resolve them on the spot if possible, and for those that were unsolvable, he would consult with his partner and then try to find a solution.

After a seminar that lasted for several hours, Xu Shenxue also had a general expectation of the manufacturing plan for the sixth-generation quantum chip.

By next spring, we should be able to completely resolve some of the remaining technical issues, build the first sixth-generation quantum chip production line, and begin various tape-out and testing processes. If the tape-out goes smoothly, the sixth-generation quantum computer is expected to enter mass production by next autumn!
The main problem is that many key technological issues in the manufacturing of sixth-generation quantum computers, especially sixth-generation quantum chips, have been solved one after another.

For example, the core equipment, the fourth-generation quantum alignment machine, has been successfully developed, even though the device itself still has many technical problems.

This fourth-generation quantum alignment machine is more of a production device than a specially developed research device. Several of its core components have a very short lifespan!

One type of high-temperature resistant material used to carry ultra-high temperature ion beams has a safe service life of only 120 hours. Exceeding this safe service life will seriously affect the processing accuracy, and continuing to force processing will cause the yield rate of quantum chips to plummet...

The yield rate of sixth-generation quantum chips processed using this equipment was already only around 20%.

If the price continues to plummet to just a few percentage points, even the exorbitant manufacturing costs will become unbearable for Zhiyun Group.

However, the fabrication of the core 3D structure of the sixth-generation quantum chip is extremely complex and requires very high precision, which means it takes a lot of time.

Ultimately, the annual production of a single fourth-generation quantum alignment machine is estimated to be only around one thousand pieces... This is the annual production, not the daily production!

In comparison, the HEUV-300C lithography machine, the top-of-the-line model currently being shipped by Gulf Technology, can produce 180 silicon wafers per hour in a single exposure, and each silicon wafer can be cut into dozens of GPU cores.

This production capacity shows that the sixth-generation superconducting quantum chip developed by Zhiyun Group is much more expensive than any semiconductor electronic chip... The manufacturing cost of each sixth-generation quantum chip is astronomical, more than ten million.

The high cost is not only due to the sixth-generation superconducting quantum chip itself, but also to the core material used extensively in superconducting quantum computers: MMK34 room-temperature and room-pressure superconductor, which is also very expensive.

So far, the secret superconducting material factory under the Zhiyun Group can only barely produce MMK34 material on a small scale. The entire production process is not much different from laboratory production, only the scale is larger... It's purely a matter of brute force, forcibly increasing production capacity through scale.

The problem of low-cost industrial mass production has not yet been solved. The 'MMK34 mass production plan' project, led by Zhiyun Group and involving a group of top domestic experts, has been underway for several years, but so far it has not achieved any major technological breakthroughs.

For a top-tier material like MMK34, which represents the pinnacle of human technology and even surpasses much of the current understanding of conventional materials science, achieving a production capacity of tens or even hundreds of thousands of tons and a price similar to that of ordinary materials like copper is extremely difficult.

Because MMK34 material cannot be mass-produced industrially at low cost for the time being, its manufacturing cost is extremely high, and this high cost will ultimately be reflected in superconducting quantum computers.

Quantum chips are expensive, MMK34 materials are expensive, and a whole host of other new technologies are also expensive... In the end, the extremely high R&D costs of the entire superconducting quantum computer industry chain must also be factored in, which results in the superconducting quantum computer under the Zhiyun Group costing hundreds of millions of yuan per unit, even for internal procurement!
However, although it is expensive, its performance is powerful enough. When used with APO graphics cards, the computing power cost of the computing center is several times cheaper than that of a computing power center composed of APO graphics cards alone.

More importantly, the significant reduction in power consumption means that more computing power can be obtained with limited power supply... This is also very important, since computing power consumes too much electricity.

If we don't use a superconducting quantum computer and instead use a traditional GPU core, the power supply requirements for a large-scale increase in computing power would be even greater... which is more difficult to solve.

Given the massive computing center construction of Zhiyun Group, if superconducting quantum computers were not used to save electricity, the domestic power grid would collapse within a few years.

After all, the electricity required by these computing centers of Zhiyun Group is staggering. Based on Zhiyun Group's three-year, 2,000 EFLPOS computing power construction goal, and the large-scale adoption of sixth-generation superconducting quantum computers and APO6000/160 graphics cards using five-nanometer technology, the estimated power consumption is approximately 500 billion kilowatt-hours…

This is already a very optimistic situation, and we still need to eliminate a large number of outdated APO graphics cards, such as the APO3000 and APO4000, which are outdated and power-consuming.

Only in this way can the power consumption be barely controlled at 500 billion kilowatt-hours!

And Zhiyun Group is not the only company in China working on artificial intelligence. Other top high-tech companies in China are also doing it, although they are not on the same scale, but each one still requires a lot of electricity.

To supply power to the numerous computing centers that have sprung up across the country, the domestic power sector is operating at full capacity, carrying out power infrastructure construction throughout the country, with various newly built power plants working day and night to put them into operation and supply power as soon as possible.

However, such important infrastructure projects take time!

For the next few years, Zhiyun Group will face a shortage of power supply... In particular, Zhiyun Group's construction of computing centers will not stop. The current plan is to build 2,000 EFLPOS, but in the future, it will build even larger-scale computing centers, such as 5,000 EFLPOS or even 10,000 EFLPOS.

By then, the power shortage problem will be even more pronounced... When Zhuyun Group communicated with the power authorities, they clearly stated:
Three years from now, the power demand of Zhuyun Group's data centers will reach 500 billion kilowatt-hours.

Five years from now, it is projected to reach one trillion kilowatt-hours... well, that's roughly equivalent to the entire annual electricity generation of Japan!
Ten years from now, the power demand of Zhiyun Group's data centers may reach two trillion kilowatt-hours, roughly equivalent to half of the electricity generated in the United States.

This timeline and the enormous power consumption have left many people speechless... Do you people at Zhiyun Group really think electricity is as easy as cabbage, something you can just get whenever you want?

Electricity issues have become one of the factors limiting Zhiyun Group's further development in the field of artificial intelligence.

Otherwise, Zhiyun Group wouldn't have gone to such great lengths to develop superconducting quantum computers, in order to obtain as much computing power as possible with limited power supply.

If a superconducting quantum computer were not used, the power demand would be several times higher...

This is also the reason why Zhiyun Group has invested heavily in the research and development of neuromorphic computer systems... The initial motivation was actually to save electricity!
The superconducting quantum computer project and the neuromorphic computer project are both crucial to the continued development of Zhiyun Group in the field of artificial intelligence, and Xu Shenxue attaches great importance to them.

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.

After a seminar was held at Gulf Technology, Xu Shenxue was notified and made a last-minute trip to a special airport to watch the maiden flight of the X3 project.

Project X is the highest level of confidentiality that Zhiyun Group is currently involved in, and it is also a core project related to its future strategy.

The entire project was officially launched in 13. After years of research and development and continuous technological iteration, the X1 prototype successfully completed its maiden flight in 17. Subsequently, the X1 project entered a continuous experimental flight phase. At the beginning of this year, it was officially and secretly finalized, and the first batch of mass-produced aircraft began to be put into service.

The X1 unmanned combat aerial vehicle (UCAV) has gradually become known to the outside world due to its frequent test flights. The information occasionally revealed in China only states that it is a stealth UAV, and the designation it uses is also somewhat misleading: Attack-18, rather than using the prefix of a fighter jet.

Previously, traditional stealth attack drones used the designation 'Attack-X'. Now, the X1 project uses the designation Attack-18. Due to habitual thinking, people still think that this is just an ordinary remote-controlled attack drone, only a bit bigger, with a maximum takeoff weight of over twenty tons...

Of course, many people suspect that this is an AI fighter jet, but it's just suspicion, and they can't find any evidence... They can't take the X1 apart to see the computing center inside!

The X1 stealth unmanned combat aircraft has now begun to enter service on a small scale, becoming an important part of the air force. Meanwhile, the X3 project, which was previously rejected and restarted, has also made the latest progress: it has ushered in its maiden flight!

After being scrapped and redesigned, the X3 project was positioned as a 'next-generation unmanned combat aircraft' with higher and more comprehensive technical performance indicators than the X1 project, and correspondingly, it also applied more cutting-edge technologies.

Now, the X3 project, after being reworked, has finally come to a conclusion... and the moment of the prototype's maiden flight has finally arrived.

As the supplier of the core 'X3 airborne artificial intelligence system', Zhiyun Group was also deeply involved, and Xu Shenxue was invited to watch the first test flight!
Looking at the newly designed X3 project in the hangar, Xu Shenxue was also somewhat shocked: This X3 drone is huge!
The X1 drone is considered quite large in the drone industry, with a maximum takeoff weight of 26 tons.

However, the X3 drone in front of us is even larger. When the staff member next to us introduced it, he said that the maximum takeoff weight of the X3 drone reaches forty tons.

This takeoff weight is heavier than that of a large number of heavy twin-engine fourth- and fifth-generation fighter jets.

For example, the F-22, according to publicly available information, has a maximum takeoff weight of between 36 and 38 tons.

The publicly stated weight of the F-15E is generally 36 tons, while that of the Su-35 is 34.5 tons.

However, the maximum takeoff weight of this X3 unmanned combat aircraft reaches forty tons.

Moreover, it is larger in size. This X3 drone also has a typical diamond-shaped tailless layout, and the flat design makes the body larger.

Moreover, according to the introduction, this plane flies exceptionally fast, capable of supersonic cruise, with a maximum speed of Mach 2.3!
Heavy, large, and fast!

This is the most obvious difference between the X3 and X1 drones. In addition to these huge differences in airframe performance, the bigger gap actually lies in the unseen internal components.

The X3 drone's onboard artificial intelligence system is even a generation ahead of the X1 drone in terms of performance!

From computing power to sensor performance, there is a gap in every aspect!
Of course, the cost difference is also significant across all aspects!
The X3 drone is very expensive!
(End of this chapter)