Military Industry of Great Powers: Build the Strongest Flanker.

Chapter 227 The Trap for Intel

Zhongxun had released the goods for two months, and the DIMM memory market prices were stable. However, in March, the memory market plummeted, and 4-megabit DIMM memory had dropped to US$.

On average, the cost price of each memory particle is already lower than that of many manufacturers.

Samsung President Lee Kun-hee always thought this was a normal contraction of the market, but after seeing the report handed over by the marketing department, he realized that something bad had happened...

"President, it has been two months and we have only shipped 70 pieces. Compared with the same period last year, the month-on-month decrease has reached %!"

This is terrible. DRAM is the main profit source of Samsung Semiconductor. Li Jianxi asked quickly: "What conclusion does your marketing department have?"

If the general environment is like this, then Samsung may want to consider investing. They are good at counter-cyclical investment and seize huge profits when the market picks up.

However, the words of Park Jimin, the president of the marketing department, gave him a hammer.

"The demand for DRAM particles in the international market has always been strong, but Kingston, HP, and Compaq have made demands on our prices. They..."

"Say!"

President Park said: "They asked for a 60% price reduction, and they were only willing to pay $512 for 8kb DRAM chips..."

Li Jianxi almost lost his breath. Even if R&D costs are not included, even if only production costs are calculated, they can't make 512kb DRAM memory particles for $8!

Li Jianxi said: "If they can tell this price, they must have found a supplier that can give them this price. What's going on?"

President Park said helplessly: "It's Chinacom, China's Chinacom, President Li. Chinacom has snatched orders from Kingston, HP, and Compaq. We found out that their 512kb DRAM memory chips only cost ten dollars, and the quality is Even better, Kingston is only willing to quote us $8 for the pellets..."

"Zhongxun!" Li Jianxi shook his head and said: "Impossible. They have just acquired Motorola's DRAM business unit. How can they launch cheap and high-performance DRAM particles? They must be dumping. They are dumping to seize the market. This is illegal. Yes, the Americans will not let them go..."

Li Jianxi suddenly had an idea, stared at President Park and said: "How will Micron react? Our costs are similar to Micron, and their products will also be affected..."

President Park shook his head and said, "So far, there has been no response from the camera!"

"..."

Dr. Andrei Gromyko, senior vice president of Micron, chief engineer and one of the company's founders, knew the news two months ago. Andrei was very confused as to why Zhongxun could make memory chips so cheap and of high quality. Good again.

As soon as Zhongxun's goods arrived at HP, their engineers got samples.

Everyone is interested in advanced technology. Micron has prepared advanced electron microscopes and high-precision grinding tools.

In the laboratory, they cut open the intestines of the memory particles, and they immediately understood why!

"Dr. Andre, just as we imagined, Zhongxun has cracked the 500nm process. With the same wafer, they can produce four times more particles than us..."

"..."

Micron's 256kb and 512kb particles still use a mature one-micron process. Unexpectedly, China has completed the commercialization of the 500nm process.

Very powerless, the commercialization of their 800-nanometer process has not yet been completed.

Even when Intel released the 486 processor the year before last, it still used a 1-micron process. It was only in the second half of last year that it completed the commercialization of the 800-nanometer process!

"Their manufacturing process is two generations ahead of ours... no wonder they make it so cheaply."

According to Moore's Law, if the manufacturing process improves one generation, twice as many transistors can be integrated into the same area, while the cost can be reduced by half.

The manufacturing process is two generations ahead. Zhongxun lowered the price of DRAM memory particles to half, and they still made a profit. However, at this price, Micron would lose three dollars if it sold one.

Laboratory engineer Bornhard Riemann shook his head and said: "More than that, Dr. Andre, they used technology we don't know about!"

Andre asked: "What technology?" Bornhard said: "Dr. Andre, come and take a look!"

In the electron microscope, Andre saw the different structures of the connecting wire metal.

"This is not aluminum!"

Bornhard said: "It's not aluminum, it's copper!"

"No wonder!" Andre shook his head and said, "The performance of copper wires is much better than that of aluminum wires. No wonder their quality is so good!"

Andre said this very weakly. The 500nm process and copper wire technology are currently the hottest topics in semiconductor research in the world. Only Intel, IBM, and Texas Instruments are spending time and money to study them, and they also have the financial resources to do so.

Unlike aluminum, copper atoms can float on the chip's insulating layer.

Copper also has the potential to alter the silicon, altering its electrical properties and disrupting the device's operability.

Another problem is that physically dislocating copper ions through ultra-fine conductors, which is commonly known as electromigration, greatly increases the probability of semiconductor failure.

It is not feasible to use a metal liner like aluminum wires and then use physical vapor deposition.

According to the original history, Intel and IBM will try the periodic table of elements over the next five years and spend billions of dollars in research and development before they find a solution for cobalt-tantalum clad copper.

In order to make Intel take some detours, Mr. Zhao added a dispensable metal deposition layer of gold and tungsten. Although the manufacturing cost increased by three US dollars per particle, it was still worth it.

Magnesium quickly used a spectrometer to analyze the materials needed for the wire.

Copper, gold, tantalum, cobalt, tungsten...

It gave me a headache. The wire actually contained five elements, not just copper. They couldn't analyze the manufacturing process.

At the same time, Intel's engineers are also digging into Zhongxun's memory particles.

They also got two technical indicators, 500nm process and copper connection.

"Their wires are about 78% copper, 12% gold, 3% cobalt and 7% tantalum?"

"What material is the substrate?"

"No substrate metal detected?"

Intel's engineers were confused at this time. There was no substrate metal? How could they solve the problem of copper deposition destroying the silicon substrate? They couldn't figure it out...

Just like aluminum and silicon base are incompatible, copper deposition is not incompatible, it can change the physical properties of silicon base. We can only add a transition layer between copper and silicon base...but...no...why ?

Why, of course I don’t want you to know.

Teacher Xiao Zhao expected that the product would be used for research, and it was also common to set several traps on the product at the same time.

"God bless Intel!"

It would be fun if they took the wrong route.

Teacher Xiao Zhao hopes Intel can be braver and more confident...

(End of this chapter)