dkrc Chapter 384 Go All Out
Chapter 384 Go All Out
On March 5, 2001, at six o'clock in the morning, the sky was not yet fully bright, and there was a layer of pale golden light on the eastern horizon, but the chill on the ground was still very strong.
When Ling Yun got out of the car, his breath dissipated in the air before his eyes. He was wearing a dark gray jacket, unbuttoned, over which he wore a black turtleneck sweater. Zhao Hu walked around from the driver's side, carrying a briefcase, and followed behind him.
The security guards at the building entrance were changing shifts. Two men in dark blue uniforms stood facing each other, one saluting, the other returning the salute. Their movements were slow and precise. As Ling Yun walked past them, both men simultaneously greeted him with "Good morning, Mr. Ling." He nodded and pushed open the door to go inside.
There were already people in the lobby on the first floor. Ni Guangnan stood next to the elevator, wearing a faded blue work uniform, the collar buttoned up tightly. He held a silver-gray metal box in his hand, about the size of a palm, very thin, like a business card holder. He was looking down at the box, his thumb rubbing back and forth on the lid, unaware that Ling Yun was walking over.
Ling Yun stood beside him without saying a word. After about ten seconds, Ni Guangnan looked up, saw Ling Yun, paused for a moment, and then handed the box over.
"The tape-out was successful, and the test results are in. All functions are normal, and the parameters basically meet our expectations."
Ling Yun took the box and opened it. Inside was a layer of black anti-static foam, with a small, silver-gray chip embedded in the center, about the size of a fingernail, its surface having a faint sheen, like freshly polished metal. He looked at it for a few seconds, then closed the box and handed it back to Ni Guangnan.
"Let's go upstairs and talk."
The two entered the elevator, with Zhao Hu following behind. Ni Guangnan pressed the button for the fifth floor. As the elevator went up, he opened the box again, glanced at it, closed it, and put it in his pocket, zipping the pocket up twice.
The two arrived at the fifth-floor office. Ni Guangnan spoke, his voice slightly hoarse, as if he hadn't slept all night, saying, "Last night at nine o'clock, the lab in Shanghai called to say the tape-out was back. We ran a round of tests, and the functionality was normal. The power consumption was within the design range. The frequency reached 200 MHz, which is not much different from the ARM7 core reference design, basically meeting our expectations."
He paused, opened the box, took out the chip, and placed it on the table. The fingernail-sized object gleamed silver-gray under the light, its surface showing circuit patterns as fine as a hair.
Ni Guangnan pressed his finger against the edge of the chip and said, "This was made on the first fully independent ARM architecture chip production line in China. The equipment was second-hand; the lithography machine was a Nikon product from ten years ago, with an accuracy of 0.35 micrometers. We modified the control system ourselves, adjusted the process parameters ourselves, and created our own testing procedures. The wafer fabrication was successful on the first try. The yield rate is 78%, which is far worse than TSMC's, but this was made entirely independently, which marks the beginning of a new era."
Ling Yun smiled at Ni Guangnan and asked, "Mr. Ni, what are the next steps? Do you have any plans for the near future?"
Ni Guangnan said, "The next step is optimization. We'll reduce the process to 0.25 micrometers, the frequency to 300 MHz, and the power consumption to 20% lower. Then we'll make system-on-a-chip (SoC), integrating the CPU, GPU, memory controller, and I/O controller onto a single chip. This is ARM's strength, and it's also our strength. We need to make breakthroughs in MCU main control, IMP, power management ICs, and other sensor areas used in home appliances. By improving our production processes, supporting equipment, and software in these areas, we can produce usable memory chips and ARM chips in another two to three years."
"Here's my idea." He pulled a folded piece of paper from his pocket, unfolded it, and laid it on the table. The paper was covered with dense writing in his own handwriting, in blue-black ink, the characters neat and legible. He pointed to the first line and read it aloud.
"The first path is to make money while developing the entire industry chain. We can't stop making money; mice and keyboards, complete computer systems, laptops, system licensing—these are all sources of cash flow. Having income every month gives us peace of mind. As for the entire industry chain, we need to be selective. We can't do everything; we don't have that much money or that many people. We need to focus on the most important things, the things we lack most. Lithography machines, etching machines, thin film deposition, and ion implantation are the four core components. Lithography machines come first; without them, nothing else matters. Etching machines come second; without them, chips can't be made. Thin film deposition comes third, and ion implantation comes fourth. Once we've got these four sorted out, we can buy the rest. If we can't buy them, we can imitate them; if we can't imitate them, we can make do. That's the priority."
He paused, his finger moving to the second paragraph on the paper. "The path to lithography machines is a three-step process. First, buy secondhand equipment, modify the control system and optical components, and reduce the process technology to 0.18 micrometers. This step will take three years. Second, collaborate with domestic optical and precision machinery research institutes to design the optical, mechanical, and control systems ourselves, and build a prototype. This will reduce the process technology to 0.13 micrometers. This step will take five years. Third, catch up with international standards: 90 nanometers, 65 nanometers, 45 nanometers. This step will take ten years. In total, it will take eighteen years. After eighteen years, we will have our own lithography machines, and we won't have to rely on others."
Ling Yun looked at Ni Guangnan and shook his head. "It's too slow. If we develop in such a step-by-step manner, we will be very passive. While we are developing, others are also progressing. The gap between us and others will not become smaller, but rather larger."
Ni Guangnan looked up at Ling Yun, "It's a bit slow, but it's steady. We take it one step at a time, without stumbling. Every step we take brings money, customers, and a market. We don't rely on government funding or bank loans; we rely on ourselves. We invest what we earn. We invest what we earn; and what we invest grows. It's a virtuous cycle. You could say we're one of the few companies in the semiconductor industry with the potential to succeed. I'm really worried that if we're not careful, this lone seedling of ours will wither away."
"Mr. Ni, time waits for no one. Once the US emerges from the dot-com bubble, the certainty of technology will increase, and a large amount of capital will flow into technology companies, especially the semiconductor and communications industries. Chip manufacturing processes will continue to shrink, gradually approaching the limits of physical theory."
Ni Guangnan nodded. "There's a second way: go all out and accelerate. Breakthroughs must be achieved simultaneously at all key nodes across the entire industry chain. Lithography machines, etching machines, thin-film deposition, and ion implantation—all four must be developed together. Optics, mechanics, materials, control, and software—all five directions must be advanced simultaneously. Within five years, prototypes of all four core pieces of equipment must be available. Don't rush to 180nm; start with 350nm. Once 350nm is done well, move on to 250nm. Once 250nm is successful, move on to 180nm. Take it one step at a time, but don't wait. Waiting one step slows you down three steps. Pushing one step fast speeds you up three steps. This path requires national policy, government funding, bank loans, and IPO financing; otherwise, even 10 billion won't last long—the entire industry chain requires far too much money."
He turned to look at Ling Yun. "We're working on four pieces of equipment simultaneously, pushing forward in five directions at the same time. The Optics Research Institute needs investment, the Precision Machinery Research Institute needs investment, the Materials Research Institute needs investment, the Control System Institute needs investment, and the Software Institute needs investment. Each direction requires at least 500 million yuan a year. Four directions, that's 2 billion yuan. Five years, 10 billion yuan. 10 billion yuan, what will we get out of that? A prototype. A prototype isn't a production line. A prototype can make samples, not products. To make products, we need to invest more money. How much? I don't know. Maybe another 10 billion yuan, maybe another 20 billion yuan. Even if we spend it all, it might not be successful. Even if it is successful, it might not be profitable. Even if it is profitable, it might not break even. It's gambling, so I don't want to go down this path."