Galactic Tech Empire

Chapter 96 Production Planning
Zhang Rujing may be too excited and a little out of breath.

Huang Haojie quickly helped him to a chair beside him.

"Haha..." Zhang Rujing laughed with tears streaming down his face: "Mr. Huang! Thank you very much for letting me join Yinhe Technology. I will die without regret. I can witness the chips of the Eastern Tang Kingdom stepping up to the top of the world. I have no regrets in this life! No regrets!"

"Dr. Zhang can still fight for Galaxy Technology for at least another ten years, and we are still far away from reaching the pinnacle of the world! This is just the pinnacle of electronic chips, and we still have quantum chips to conquer." Huang Haojie said while giving Zhang Rujing Shun Shun back.

"That's right, I still want to witness the bright moon above Mount Everest." Zhang Rujing calmed down a little.

"Du Wei!" Huang Haojie shouted loudly.

"Boss Huang!" Du Wei hurriedly stood up.

"Now your task is the most important. Among the three chariots of the Galaxy Group, Neutron Star has completed a large silicon wafer, and Galaxy Semiconductor has successfully built a 5-nanometer chip production line. Now it depends on the design of your Star Map. "

"I will do my best so that Xingtu Company will not hold back." Du Wei also felt a lot of pressure.

"In this way, you will design chips together with Xinghe Semiconductor from now on, don't be afraid to spend money, tape out while designing, and keep trying. I don't believe that you can't improve your chip design capabilities." Huang Haojie ordered firmly.

"Yes, our Xingtu company will definitely live up to expectations." Of course Du Wei knew that if he did this, his chip design ability would improve the fastest.

After all, talking on paper on the computer has no real effect on the production line. Practice is the only criterion for testing the truth.

This is also the only way for Xingtu Company to improve its design capabilities to the world level, that is to keep trying and tape out.

Because on the design diagram of the computer, you can never know whether the circuit diagram is good or bad, and you don't know which ones are good designs and which ones are defective designs.

This is why in semiconductor production, the strong are always strong, such as Taijidian and Sanxin, they can continuously find design defects on the production line, so as to achieve precise improvement.

And those small factories can't afford this kind of thing at all.

Tape-out is to manufacture chips through a series of process steps like an assembly line, which is tape-out.

In the field of integrated circuit design, "tape-out" refers to "trial production", which means that after the circuit is designed, a few or dozens of pieces are produced for testing.

If the test passes, mass production will start like this.

In the chip manufacturing process, there are generally two periods of time that can be called tape-out.

The chip design company finds that a certain place can be modified to achieve better results when designing, but is afraid that such modification will bring unexpected consequences to the chip. If a large-scale chip is manufactured based on such a problematic design , then the loss will be large.

Therefore, in order to test whether the integrated circuit design is successful, it is necessary to carry out tape-out, that is, from a circuit diagram to a chip, to check whether each process step is feasible, and to check whether the circuit has the performance and functions we want.

If the tape-out is successful, chips can be manufactured on a large scale; otherwise, we need to find out the reasons and optimize the design accordingly.

Why taping out is very expensive at the beginning is because there are many processes that need to be verified at the beginning and the production volume is very small. For a good chip, the traditional production method generally needs to go through hundreds of processes, the production cycle is long, and the cost of taping out is almost The cost is tens of millions or even hundreds of millions.

The reliability of chip manufacturing equipment is very important. After the development of new equipment, it is still a long way to wait for customer certification, large-scale production and mass promotion.

Therefore, the research and development unit invests a lot in the early stage.

When tape-out is mass-produced, all process verification has been completed, all costs are effectively controlled, and large-scale production will bring better economic benefits to the production unit.

Therefore, large-scale tape-out production is relatively cheaper.

A tap-out of a traditional chip costs tens of millions to hundreds of millions. Who dares to stream a chip for no reason? Even Taijidian and Sanxin are making adjustments while producing.

Using other people's money for trial production, of course, you can continue to try. If it is a chip factory that has no orders, you let him trial production for no reason, and it will go bankrupt in minutes.

Therefore, semiconductors do not have tens of billions of dollars of investment, just think about it.

As far as Galaxy Technology is concerned, the price of trial production tape-out is very low.

First, the raw materials for the production of chips are all produced by ourselves.Second, Galaxy Semiconductor's process production cost is very low, only equivalent to about one-tenth of traditional chips.

In general, how much does traditional chip production cost?
It is very simple, the larger the scale, the lower the cost, and the smaller the chip manufacturing process, the lower the cost. If the chip manufacturing process is 16 nanometers, 100 million pieces are produced, the cost is 30 US dollars; then the production process is 7 nanometers, and 1000 million pieces are produced, and the production cost is 20 dollars.

If it is to produce hundreds of millions of pieces, the cost can be reduced to 10 dollars, or even lower.

According to the 8:20 pricing method of the international low-profit chip design company's pricing strategy, that is, when the hardware cost is 8, the price is 20.

Don’t think this price is high, it’s already very low. Intel’s general pricing strategy is 8:35, and AMD has historically reached 8:50.

The Qualcomm Snapdragon 845, which just came out this year, is priced at around US$70, which means that the cost of the Qualcomm Snapdragon 845 is around US$15.

Galaxy Semiconductor can keep costs down, on the one hand, because raw materials such as silicon wafers are produced by itself;
On the other hand, because the new process does not require the use of auxiliary materials such as photoresist, these auxiliary materials account for about 30% of the cost of chip production.

At the same time, because the production speed of the new process is very fast, this leads to shortened production time, which means that the unit labor cost and electricity cost are reduced.

After all, for lithography machines and the like, if those lasers are used for a long time, the electricity bill alone is enough.

The unit cost of Galaxy Semiconductor is 20-30 Wah, which is equivalent to 3-4 US dollars.

Of course, this is without calculating the depreciation cost of the equipment. Galaxy Semiconductor built such a production line, which can produce 2300 million chips a year. It cost a total of more than 5000 million Chinese yuan. If the depreciation cost of the production line is evenly spread, it is equivalent to the increase of each chip. Two bucks cost.

In fact, the really expensive atomic imprinting machine, the Galaxy Semiconductor people like to call it the chip mother machine, the cost of one set is about 1.2 million Chinese yuan.

In order to enhance the strength of Galaxy Semiconductor, 20 atomic imprinting machines were built, and a total of 24 billion Huayuan was distributed.

However, if these are produced on a large scale, the cost will become lower and lower.

Therefore, the tape-out of Xinghe Semiconductor is about [-] yuan at a time. Huang Haojie does not believe that Xingtu Company has not made any progress.

Next, Huang Haojie has already discussed with Wang Bosi, Zhang Rujing and others, and let Xingtu Company design and streamline improvements at the same time.

At the same time, before March 2018, the scale of the chip production line will be increased to 3.

The other is research and development of packaging and packaging and testing equipment, because the strongest and largest packaging and testing company in the world is Dongdao ASE, plus other packaging and testing companies in Dongdao, this field is monopolized by them.

Don't underestimate the packaging, testing and packaging. These two links sometimes cost a lot.

Packaging is to stack the substrate, core, and heat sink together to form the CPU that everyone sees every day. The packaging cost is the funds required for this process.

Under normal circumstances with a huge output, the packaging cost generally accounts for about 5%-25% of the hardware cost, but some of IBM's chip packaging costs account for about half of the total cost, and it is said that the highest has reached 70%.

The test can identify the key characteristics of each processor, such as the highest frequency, power consumption, heat generation, etc., and determine the level of the processor, such as classifying a bunch of chips into categories: I5 4460, I5 4590, I5 4690, I5 4690K Wait, after that Intel can offer different prices according to different grades.

However, if the chip yield is large enough, the test cost can be ignored.

Therefore, Huang Haojie does not intend to let other companies package and test chips, because Galaxy Semiconductor can do it by itself, and there is no need for others to make money.

What's more, the packaging, packaging and testing companies in China are not good. Is it necessary to send them across the ocean to Dongdao for packaging, packaging and testing? How much will the cost of this round-trip increase?
Instead of wasting money and time looking for others to package and test, it is better to be self-reliant.

(End of this chapter)