The flowers bloom at two ends, one on each branch.

What happened to Wang An's computer in Silicon Valley, of course Gu Biao didn't know.

The inside stories that Chambers and Caningham knew could not have been directly revealed to them by Gu Biao.

Gu Wei has no interest in the secrets of the real national science and technology strategy, and what he inadvertently reveals to Peter Peterson here is nothing more than some false information taken out of context.

From those pictures and voices that pinch the head and tail off, the Americans can at best come to the conclusion that:

"The Chinese really love peace and are willing to support the American-style global division of labor. I already know that it is hopeless to independently develop an open personal computer operating system to compete with the United States, so I obediently learn to deposit the money in the bank to eat interest and wait for the advanced goods to appear."

As for what happened in the next link, there is no need to repeat it.

At the same time that Wang An's computer was in turmoil, Gu Biao had just arrived in Xiangjiang and visited the front line to inspect the production of Xiangji and Hanle Electronics. (Hanle Electronics is producing the newly developed "Tiankun Learning Machine")

The first thing he went to was Xiangjidian.

Zhang Zhongmou returned to Xiangjiang a few days earlier than Gu Biao, so when Gu Biao went, it was Zhang Zhongmou who personally received him.

After this trip to the capital, and even the various resources that Gu Biao coordinated for the smooth launch of Xiangjidian, Zhang Zhongmou's admiration for Gu Biao this year has risen to a new level, and his respect for Gu Biao's strategic vision has also reached the level of obedience.

I think that before October last year, when Gu Biao behaved like a black-hearted person with money and no discipline, Zhang Zhongmou once regarded Gu Biao as a person who was "fierce and timid, and had no foresight".

And now, he has recognized the reality: Gu Biao is the Dinghai Shenshen Needle for the smooth development of Xiangji, from the blockade of equipment to the lack of early customers and early collaborators, all rely on Gu Biao to push the level.

Zhang Zhongmou figured out for himself, if there was no Gu Biao and completely relied on himself, the development speed of Xiangji would be at least two years slower - this estimate is very accurate, because TSMC in another time and space, the horizontal comparison of the development speed is at least two years slower, and that is still with the full support of the Wanwan Academy Development Fund.

Since Zhang Zhongmou is like this, the other management of Xiangji Electronics is of course also a five-body investment for Gu Biao, a big man who rarely shows up but is said to hold the absolute largest shares.

There will be no hidden dangers of internal unity and prestige for at least a few years.

……

"This is the latest generation of Motorola 68000 CPU manufactured by us, 6MHZ main frequency CPU, the current main product. After our previous struggle and running-in, the current comprehensive processing cost is 10%~15% lower than that of Toshiba and NEC, and there is no advantage over NEC in terms of accuracy and yield.

However, if it is compared with Texas Instruments, our comprehensive processing cost can be reduced by 20%, and the yield rate is already higher than that of Texas Instruments. ”

Outside the dust-free workshop, a production line engineer wearing a mask took out a plate of samples that had just come off the production line for Gu Biao to review, and there was a microscope for visual inspection next to him. And Zhang Zhongmou stood aside to explain.

The inspection link is set up very professionally, and next to the visual inspection microscope, there are also Motorola's self-produced goods last year when it relied on Deyi outsourcing, as well as the control OEM products of this NEC. And each batch has hundreds of pieces, in order to make it easy to see the yield rate and process stability.

After all, just take out one or two pieces of the best quality, that is not a problem, the key is a large number of random sampling.

In 1986, the chip foundry industry was still relatively mixed, and there was no such thing as "chip lithography and packaging can still have process defects?" as later generations of consumers thought. X or more?"

This is the result of TSMC and other companies that have worked hard for ten or twenty years for quality control, and it is not innate.

At this point, I have to admit that the Wanwan people in the later world do have a set of chip lithography/packaging quality control, which is almost on par with Yue himself, and the quality control of Koreans is not as good as that of Yueben and Wanwan.

Gu Biao didn't know much about the technical details in this area, but at this moment, under the guidance of the production line engineer, he knew how to look at the quality of the product and the level of the process.

After sampling for a while, Gu Biao rubbed the bridge of his sore nose, removed his glasses from the microscope for visual inspection, and asked:

"What is the cost of a 68,000 piece that we are producing at the moment, what is the price point for Motorola, and what if it is for our own use?"

Zhang Zhongmou couldn't hear the joy and anger of the explanation: "The semiconductors are still imported from my side, and the cost of all materials is about 40 US dollars, and our settlement to Motorola is 82 US dollars, and Motorola's official guide price to the outside world is 116 US dollars."

In addition, if we directly deduct the technology license fee for the production of the Tiankun learning machine, Motorola will give us a license price of $96, which is equivalent to a technical license fee of only $14 for us, which is $20 cheaper than the outside world. ”

Gu Biao replied and asked: "Our Maori 42 and Motorola Maori 34? How do I remember that when we first negotiated, Motorola's net profit per chip was much higher than ours. ”

Zhang Zhongmou smiled bitterly: "It's much higher than us - this $42 is not gross profit, and the depreciation of the equipment has not been deducted." So many lithography machines, packaging and supporting equipment. The things bought for 400 million dollars are depreciated in about seven or eight years. There are still 300 million yuan of equipment in the follow-up, which will be in place one after another, and it will also be amortized.

I calculated, this year's annual production capacity is 5 million pieces of 68000CPU, which is relatively low, after all, the trial flow will only start in March, and the scale of the factory will be completed in stages. The depreciation of equipment in that year was 50 million US dollars, which is equivalent to 10 US dollars for every CPU produced.

In addition, this year's total salary and labor costs will also be 50 million US dollars, and it will also cost 10 US dollars to amortize each CPU. At present, we have more than 1,600 employees, and the salary in Xiangjiang is high, even if you find a production line worker in a dust-free workshop, the total cost of a year will add up to 10,000 US dollars. A production line engineer with education and experience may start at $30,000.

At present, there are about 1,200 front-line workers, production line engineers, and management positions, and the total labor cost is 20 million US dollars per year. There are also 400 R&D personnel, the annual cost is 30 million US dollars - but these people are necessary, before you engaged in Toshiba, it only shortened the gap between us and myself in the initial stage, but then we want to surpass myself, I calculated, this R & D investment is absolutely not lower. ”

According to this calculation, the depreciation of the equipment amortizes the cost of $10 per piece, and the labor also spends $10 per piece (workers and management only account for $4, and the majority is $6 of R&D amortization), leaving the gross profit per piece to only $22.

The $22 also deducts the depreciation of the plant, the rent of the land, and the financial cost...... All of them are deducted, and every 68,000 CPU produced is a good net profit of $15.

However, Gu Biao asked Xiangji to directly supply Tiankun Entertainment according to the "price of supplying Motorola" of 82 US dollars, so in this transaction, Xiangji lost 14 US dollars in licensing fees. (The $14 is the patent fee to be given to Motorola, because the 68000 CPU was developed and designed by Motorola, and the OEM company has to intercept the current for its own use after it comes out, and the licensing fee still has to be paid)

In this way, Xiangji is completely a day job, using OEM profits to fill the licensing fee, and there is almost no profit this year.

It is precisely because of this that Gu Biao's Tiankun learning machine can still only sell for a total price of $199 when the CPU and peripherals are upgraded compared to the prototype game console. Otherwise, if you enter the CPU at $96, Tiankun will lose money.

If it were another person who did this, this kind of thing would definitely be a "related party transaction", which is a typical example of "the interests of major shareholders are transferred to affiliated companies and infringe on the interests of the company's minority shareholders" in the Company Law. (The gross profit of Xiangji was transferred to Tiankun, because Gu Biao's shareholding in Tiankun was much higher than that in Xiangji)

However, if Gu Biao does this, it will be fine for the time being.

Because Xiangji was originally a company that had just been put into production this year, your first big customer was pulled by Gu Biao. People give you the first drop of blood to continue the order, you don't make money to work as a day job, isn't it right? There is a kind of waiting for your wings to harden, you don't have to worry about giving resources, you can also pull enough real ability orders.

Zhang Zhongmou and other small shareholders also acquiesced to Gu Biao's operation.

This means that if you want to really make money in a year or two, you have to count on more production capacity and sales, and find a way to find a way to get rid of the "average $20 per CPU amortization" cost of equipment depreciation and human resources.

diluted that part of the money, Zhang Zhongmou can still make money.

Therefore, Gu Biao naturally asked: "As you just said, because the factory has not yet been fully completed, and it only started in March this year, the production capacity is low." So, to what extent can you expand your production capacity next year and next? Are you sure that the market sales volume will keep up, and how much will the amortization cost rise?"

ZHANG ZHONGMOU THOUGHT FOR A WHILE: "IN THE PAST FEW YEARS, THE PRODUCTION CAPACITY MUST HAVE BEEN STEADILY INCREASED, AS YOU SAID, IN FISCAL YEAR 1987, I ESTIMATE THAT IT CAN REACH AN ANNUAL OUTPUT OF 10 MILLION PIECES OF 68,000-6MHZ PROCESSING PRECISION PRODUCTS, 12 MILLION PIECES IN FISCAL YEAR 88, AND 8 MILLION PIECES OF NEXT-GENERATION 10HMZ PROCESSING ACCURACY CHIPS, AND THEN GRADUALLY ROTATE AND ELIMINATE TO EXPAND PRODUCTION CAPACITY."

By the end of FY88, annual equipment depreciation will also double from today to $100 million per year. However, labor costs will not increase significantly, and at most they will be fine-tuned according to the average annual wage increase. After 90 years, the depreciation rate of equipment will stabilize.

In other words, within two years, I was able to reduce the depreciation of equipment from $10 to $5 per CPU machined, and then I could reduce it slightly. If the market demand can keep up, we ensure that 365 days of two shifts or three shifts, so that the equipment is not idle, this piece still has potential - you know, the reason why the equipment is set for 6~8 years of rotation is not bad, but because of Moore's Law, it will lag behind in this year, and even small profits can not be done. ”