eurekapd Chapter 197 Zhang Sheng
Chapter 197 Zhang Sheng
On the last day of February, Hangzhou experienced a moderate rain.
In the office area of Vilan Microsystems, Zhao Guoping was checking the logistics information for the last few shipments. 150,800 sensors were divided into nine batches, with the farthest batch destined for Chengdu Zhenxin Microelectronics via SF Express air freight, expected to arrive on the morning of March 1st.
"Mr. Zhao, Huarui has signed for it." Xiao Chen from the quality inspection team walked over with his phone, the screen showing a screenshot of the logistics confirmation.
Zhao Guoping nodded and ticked a box on the form. Five of the nine companies had already signed for delivery, and the remaining four would all arrive by tomorrow at the latest. He glanced at the gray sky outside the window, but felt a sense of relief—this was the first official commercial delivery since the establishment of Weilan, 150,000 sensors, each of which had undergone a pre-inspection process that he had personally signed off on.
But what Zhao Guoping didn't know was that on the very morning he was checking the logistics, the online discussion about Wei Lan was undergoing a subtle shift.
Half a month has passed since He Wentao's lengthy article was published in Semiconductor Industry Observer, but its influence has not diminished at all. On the contrary, like a pebble thrown into a pond, the ripples continue to spread.
On February 28, Professor Liu Jianping of the Department of Microelectronics at Tsinghua University published a short article of 3,000 words on his WeChat public account, titled "Rationality and Frenzy in the MEMS Industry from the Perspective of a Third-Order Model".
The core argument of the article is simple: He Wentao is right.
Liu Jianping listed six MEMS cases from the past twenty years that "achieved theoretical breakthroughs but failed in industrialization," from the flexible MEMS from the Fraunhofer Institute in Germany in 2003 to the graphene-based sensor from MIT in 2018. Each case had a common feature—the papers were impressive and the data was impressive, but they all collapsed when it came to mass production.
"Weilan's third-order model is indeed theoretically innovative," Liu Jianping wrote at the end of the article, "but we must acknowledge a harsh reality: Nature Microtechnology's rejection rate exceeds 90%. Even if a paper is eventually accepted, the gap between publication and mass production has already swallowed up too many equally excellent teams. The industry should remain rational until it sees mass production data verified by independent third parties."
This article has had a greater impact on the engineering community than He Wentao's. The reason is simple: Liu Jianping is from Tsinghua University, and He Wentao is from Fudan University. The fact that two MEMS scholars from top domestic universities are simultaneously expressing similar views carries a signal significance in the industry far exceeding the articles themselves.
That afternoon, a reprinted article on Huxiu changed the title to a more widely disseminated version: "Two professors from Tsinghua and Fudan University simultaneously warn: Wei Lan's third-order model may just be theoretical."
The title may be sensationalist, but the effect is real.
This article garnered over 80,000 views on Huxiu, and the sentiment in the comments section has shifted from "wait and see with a hint of optimism" two weeks ago to "wait and see with a hint of pessimism."
"Ultimately, it all comes down to the thesis. If the thesis doesn't pass, everything else is for nothing."
"NM has a 90% rejection rate. Can a startup like Vilan really get accepted? I think it's doubtful."
"The failure examples He Wentao mentioned are truly disheartening, especially the one in Japan in 2008. Back then, they were being touted all over the world, but what happened?"
"Objectively speaking, Su Chen's data does look good, but without independent verification, there is no independent verification, and that's the bottom line of science."
The situation on Zhihu is similar. The anonymous user known as "Calm Analyst" updated their status again after two weeks of silence, this time with a noticeably more cautious tone:
Regarding the latest assessment of Weilan: I largely agree with Professor Liu Jianping's analysis. I remain optimistic about the theoretical consistency of the third-order model, but the uncertainty surrounding its industrialization is indeed increasing. Currently, there is only one key variable—the peer review results from NM. If the paper is rejected, Weilan will face a very passive situation. If the paper is accepted, even with major revision, it will be a positive sign. Everyone is waiting.
This post received 2,800 likes and ranked first in the "Vilan Micro System" topic.
And "Semiconductor Veteran 2003," who hadn't spoken all along, remained silent.
Su Chen saw all of this, of course.
On the morning of March 1st, he sat in front of the computer in the lab and read Liu Jianping's article from beginning to end. Then he closed the browser and opened his own data analysis folder.
Fifteen sets of 400mm edge verification data were neatly arranged on the screen. R²=0.978, S-shaped turning position deviation less than 0.3 mm, T1-T2 deviation 1.3%. He had seen these numbers countless times, but every time he looked at them, he could feel a steady sense of certainty.
Liu Jianping is both right and wrong.
The correct part is: indeed, too many teams have failed at the hurdle between publishing papers and mass production. The incorrect part is: he didn't know that Wei Lan had already crossed that hurdle. The 150,000 sensors weren't lab samples; they were official commercial products from CR Microelectronics' mass production line, with a yield rate of 93.6%.
But Su Chen didn't intend to say anything.
It wasn't because he didn't care about external opinions, but because he knew very well that any verbal statement was meaningless before the paper was published. The academic community values data and peer review, not press releases.
He sent Zhou Zhiyuan a message: "Have you read Liu Jianping's article?"
Zhou Zhiyuan replied quickly: "I've read it. I'm familiar with all six failure cases he listed; I even wrote commentary articles on three of them back then. His analysis is sound, but his conclusion is based on a flawed premise—he thinks we're still in the laboratory stage."
Su Chen typed two words: "No response."
"Of course we won't respond," Zhou Zhiyuan said. "Let them wait. The peer review comments should be available soon, at the latest by early April. By then, everyone will be shut up."
Su Chen hummed in agreement, then closed the chat window and continued with his work.
What he is doing is building a data analysis framework for his second paper. If the core of the first paper was the theoretical derivation of the third-order model and the experimental verification at 300mm, then the core of the second paper is the experimental evidence of the behavior at the 400mm edge—fifteen sets of S-shaped turning point data, statistical significance of R²=0.978, and most importantly: proof of cross-scale predictive ability.
He and Zhou Zhiyuan had already agreed on the target journal for this paper: Advanced Materials. It's not a Nature sub-journal, but its influence in the field of materials science is equally top-notch. Moreover, the review cycle for AM is usually shorter than that for NM. If the revision of the first paper goes smoothly, the second paper can be submitted immediately after the first one is formally published, creating a series of successes.
He never uttered the phrase "defense is the best offense" to anyone, but it was indeed his strategy. His first paper was defensive—establishing academic legitimacy through peer review. His second paper was offensive—using 400mm data to prove that the third-order model was not only theoretically consistent but also possessed cross-scale predictive capabilities and practical mass production value.
Outside of the academic track, progress on the product line is also proceeding steadily.
At 3 p.m. on March 1st, Zhao Guoping's walkie-talkie rang.
"Mr. Zhao! Zhenxin has been signed for! The last one!"
Zhao Guoping stood up, walked to the window, and looked at the sky after the rain. All nine companies had signed for the delivery. The 150,800 sensors, from the first batch of wafers rolling off the production line on January 27th to all shipments on February 26th, and finally delivery on March 1st, took a total of 33 days.
He picked up his phone and sent Lin Wei a message: "All nine stores have signed for delivery. First batch of commercial deliveries complete."
Lin Wei replied with a simple "Received," followed by, "Next comes the most crucial step—customer integration testing. Keep a close eye on Zhenxin; their autonomous driving inertial navigation module has the most complex application scenario."
Zhao Guoping understood Lin Wei's meaning. Sending out the sensors was only the first step; the real test was whether customers could successfully integrate them into their own products and whether the performance data after integration would meet the standards.
If the test data of Zhenxin's inertial navigation module looks good, it will be a practical demonstration of Vilan Sensor's performance in the most demanding application scenarios, which is more convincing than any academic paper.
On the same day that Vilan was quietly pushing forward with product delivery, three thousand kilometers away in Stuttgart, a major event occurred at Bosch headquarters that drew the attention of the global MEMS industry.
At 10:00 AM Central European Time on March 1st, Bosch Semiconductor released a 42-page strategic white paper on its official website, titled "MEMS 2025: A Technology Roadmap for Next-Generation Sensor Production Lines".
The core message of the white paper can be summarized in one sentence: Bosch will invest 180 million euros over the next three years to upgrade its main MEMS factory in Reutlingen from a 300mm production line to a 400mm production line, with the goal of achieving mass production of 400mm MEMS sensors in the fourth quarter of 2024.
180 million euros.
What does this figure mean in the MEMS industry? From its inception to the present, Vilan Microsystems' total R&D investment, equipment procurement, and personnel salaries have amounted to approximately 30 million RMB. Bosch's investment this time is nearly forty times Vilan's total historical investment.
The white paper was released by Klaus Albrecht, President of Bosch's Semiconductor division. In a video statement released alongside the white paper, Albrecht, wearing his signature navy blue suit, addressed the camera in a calm tone:
"There is no doubt that 400mm MEMS is the future of the industry. As a global leader in the MEMS industry, Bosch has both the responsibility and the capability to take the lead in achieving this technological leap. We have already achieved satisfactory initial results in the laboratory stage, and now it is time to bring these results to industrialization."
This statement was simultaneously reprinted by Reuters, Bloomberg, and Nikkei Asia, triggering a chain reaction in global semiconductor industry media.
Meanwhile, on the other side of the ocean, at almost the same time, another piece of news caused a stir in the industry.
At 2 PM Beijing time on March 1st, STMicroelectronics' official Twitter account posted a tweet:
"STMicroelectronics is pleased to announce a joint research initiative with Kyoto University's Advanced MEMS Laboratory on next-generation 400mm MEMS thermal-elastic coupling technology. This multi-year collaboration aims to bridge the gap between theoretical models and industrial-scale manufacturing."
STMicroelectronics announced a joint research project with the Advanced MEMS Laboratory at Kyoto University, focusing on 400mm MEMS thermoelastic coupling technology.
Thermoelastic coupling.
The implications of those four words are crystal clear. Wei Lan's third-order model is officially called the "Third-Order Nonlinear Extended Model of MEMS Thermoelastic Coupling." Although STMicroelectronics' tweet didn't explicitly mention Wei Lan's name, any person working in the MEMS industry can understand its meaning:
STMicroelectronics wants to bypass Vilan and manufacture its own 400mm chips.
Moreover, their partner is Kyoto University—one of the strongest centers for MEMS research in Japan. Kyoto University's Yamamoto Laboratory has accumulated experience in MEMS thermodynamics second only to the Vogt team at the Technical University of Munich, ranking among the top five research groups globally.
Two announcements were released on the same day: Bosch and STMicroelectronics—the world’s first and third largest MEMS companies—simultaneously announced their entry into the 400mm market.
This has caused a significant shock within the industry.
Semiconductor Industry Observer published a quick commentary that afternoon: "Bosch bets €1.8 million on 400mm, ST-France joins forces with Kyoto University—How long can Vilan maintain its first-mover advantage?"
The article's core argument is that while Vilan's third-order model did open the theoretical door to 400mm MEMS, its first-mover advantage may soon be caught up with or even surpassed once players like Bosch and STMicroelectronics enter the market. The reason is simple: Vilan lacks capital, production lines, and a global supply chain. Bosch and STMicroelectronics have all of these.
"It's like someone discovering a gold mine, but all he has is a shovel. Now, a large company has arrived with excavators," the article used this analogy.
This article spread rapidly within the industry. By evening, uneasy voices began to emerge in the Vilan Alliance's WeChat group.
Mr. Wang, the general manager of Jingce Microelectronics, was the first to post a message: "Has everyone seen Bosch's white paper? 1.8 million euros!"
No one replied.
Five minutes later, Mr. Li, the manager of AAC Technologies' MEMS business unit, said, "We are discussing this internally."
Still no one has replied.
Lin Wei saw the messages, but she didn't say anything in the group chat. She dialed Zhao Guoping's number: "Has the timeline for the first batch of customer feedback been finalized?"
"Zhenxin said they need seven to ten days to complete the integration testing," Zhao Guoping said. "The first batch of data should be available around March 8th to 10th."
"Hurry up. The sooner the better." Lin Wei's voice was calm, but Zhao Guoping could hear the urgency in it.
After hanging up the phone, Lin Wei sat at her desk, looking out the window at the darkening sky.
Bosch invests €180 million in a Franco-Italian collaboration with Kyoto University.
If these two pieces of news had been released three months ago, she might have panicked. But not now. Because she has something that neither Bosch nor STMicroelectronics has—150,000 mass-produced sensors already delivered to customers, and upcoming real-world test data.
Bosch's white paper says mass production will begin in Q4 2024? Vilan's sensors are already in the hands of customers. STMicroelectronics is going to conduct joint research with Kyoto University? Vilan has already obtained R²=0.978 for fifteen sets of 400mm data.
There is only one question: when will these facts be made known to everyone?
Lin Wei picked up her phone and sent Su Chen a message: "Did you see what happened with Bosch and STMicroelectronics?"
Su Chen's reply was brief: "I saw it. As expected."
What do you think?
"They're proving us right," Su Chen typed, then added, "If 400mm had no value, Bosch wouldn't have spent 1.8 million euros."
Lin Wei stared at the two lines of text for a few seconds, then the corners of her mouth curled up slightly.
Su Chen is right. Bosch and STMicroelectronics' actions, on the surface, appear to be a competitive threat, but from another perspective, they are precisely the strongest endorsement of Vilan's third-order model—the world's number one and number three MEMS giants are voting with real money, proving that Vilan's chosen direction is correct.
But not everyone can see this truth.
The silence in the alliance group lasted all night. The next day, March 2nd, Mr. Wang from Jingce Microelectronics privately arranged to meet Lin Wei.
The meeting took place at a teahouse by West Lake in Hangzhou. Mr. Wang got straight to the point: "Mr. Lin, to be frank, Bosch contacted us."
Lin Wei did not move her hand holding the teacup.
"It's not a formal collaboration," Mr. Wang quickly added. "It's their Asia-Pacific market team conducting research to understand the domestic MEMS supply chain. But they asked some very detailed questions, such as our cooperation model with Vilan, our technology integration methods, and production capacity planning."
Lin Wei put down her teacup: "So you want to bet on both lines."
Mr. Wang did not deny it: "I have no intention of withdrawing from the alliance. But as you understand, Jingce is a listed company, and I need to be responsible to its shareholders. Bosch's investment of 1.8 million euros is not a small amount. If their 400mm production line really goes into mass production in 2024, we can't put all our eggs in one basket."
Lin Wei looked at him and remained silent for a few seconds.
"Mr. Wang, I understand your considerations," she said calmly, "but I suggest you wait a week."
"What are you waiting for?"
"Wait for a set of data."
Mr. Wang frowned: "What data?"
Lin Wei did not answer directly: "A week later, you will see something that Bosch couldn't buy even for 1.8 million euros."
Mr. Wang frowned even deeper, but he finally nodded: "Okay, one week."
Lin Wei stood up, smiled slightly at him, and then turned and left the teahouse.
As Lin Wei stepped out of the teahouse, the early March sun shone brightly in Hangzhou. She glanced at her phone; the integration testing at Zhenxin was underway.
Seven days.
She needs seven days.