Tech giants start by disassembling systems

Chapter 204 Encirclement

April 18th.

A conversation took place between people in three different time zones, without a meeting room, an agenda, or any written record.

The initiative was initiated by Stein, from Bosch's sensor division.

The participants were Marco Bertoli, head of the MEMS business unit at STMicroelectronics, and Thomas Weber, head of the sensor product line at Infineon.

The format was an encrypted video call. The time was 9:00 AM Central European Time—working time for Stein and Bertoli, and 4:00 PM for Weber, who was on a business trip in Tokyo.

There is no formal cooperative relationship between these three individuals. In fact, Bosch, STMicroelectronics, and Infineon are the most direct competitors in the MEMS sensor market—the three companies together account for more than 60% of the global market share.

But today, they have a common topic.

"Thank you both for taking the time." Stein spoke calmly, as if discussing a quarterly report. "I'd like to share some internal evaluation findings regarding the Villan third-order model and hear your thoughts."

"We're doing an assessment too," Bertoli said. His voice, with a slight Italian accent, came from his Milan office. "Please continue."

Weber didn't speak, only nodded slightly. He sat in the business center of a Tokyo hotel, his tie loose, shirt cuffs rolled up, looking like he had just finished a long meeting.

"There are three core conclusions." Stein opened his laptop, but didn't share the screen—it's customary for such informal conversations to leave no electronic traces. "First, the third-order model has been independently verified in terms of theoretical accuracy by multiple parties. Our initial internal reproduction also supports this conclusion. At a 400mm scale, the prediction accuracy of the third-order model is four to six times higher than our existing second-order model."

"Our conclusions are similar," Bertoli said. "I had Kosuke Yamamoto's team perform calculations using data from our joint research with Kyoto University. The results are highly consistent with the data published by Vilan."

"Secondly," Stein continued, "the NM paper on the third-order model is currently in the Major Revision stage and will most likely be accepted. According to our information, their revision has already been submitted, which is exceptionally fast."

"How did you know?" Weber finally spoke. His voice was a little hoarse—the fatigue from the business trip to Tokyo was written all over his face.

"Academic channels." Stein didn't elaborate. "Although NM's peer review process is confidential, the academic circle is very small."

Weber nodded and did not ask any further questions.

"Third," Stein said, "and most importantly—if the third-order model is formally published by NM, it is very likely to become the de facto standard for MEMS thermoelastic coupling modeling within two to three years. All three of our 400mm production lines will require this theoretical framework."

There was a few seconds of silence in the call.

"Stein, what do you mean?" Bertoli asked.

"What I mean is, we need to discuss our respective response strategies. We need to avoid duplication of effort and also avoid being passive."

……

Bertolli first gave the Italian assessment.

"Frankly, I don't think Vilan poses a direct market threat in the short term." He spoke confidently—confident because STMicroelectronics holds a 17% share of the global MEMS market. "Their monthly production capacity is around 20,000 to 30,000 units. Our daily shipments at STMicroelectronics exceed that number. The production capacity difference is more than two orders of magnitude."

"Production capacity is not something I'm worried about," Stein said.

"I know you're worried about theory," Bertoli nodded, "but there's a long way to go between theory and industry. Villand has the theory, but they don't have a complete supply chain, a global sales network, or sufficient production capacity to support large-scale commercialization. They'll need at least three to five years to become our direct competitor—if they can hold out that long."

"'If they can hold out that long'—what do you mean by that?" Stein's tone changed slightly.

Bertoli was silent for a moment.

"What I mean is, a startup will encounter countless obstacles on its way from a theoretical breakthrough to becoming a major player in the industry. Funding, talent, supply chain stability, customer trust... many startups fail along the way. We don't need to do anything proactively; time itself is our ally."

"I disagree," Stein said calmly but firmly. "Marco, three months ago, Vilan only had lab data. Now they have triple independent validation, community reproduction, a Nature News report, and a paper about to be accepted by NM. In three months, they've accomplished what many teams take three years to achieve. If we rely on 'time' to solve problems, we might find that time is on their side."

Another silence.

Weber finally joined the discussion. His voice was clearer than before—the topic had clearly caught his attention.

"I agree with Stein's assessment. The progress at Vilan has indeed exceeded our expectations. We at Infineon initiated a special assessment last week." He paused, "But I also understand Marco's point. In the short term, the production capacity gap is indeed a natural barrier. The question is whether we should proactively reinforce this barrier."

"What do you mean?" Stein asked.

"What I mean is—" Weber carefully chose his words, "should we take proactive measures in some areas instead of passively waiting?"

Bertoli had clearly been waiting for this topic. He immediately took over the conversation.

"I have some ideas."

His voice became more composed—a tone he used when discussing strategy, quite different from his previous casual manner when discussing technology.

"The first direction is at the academic level. The third-order model has received a lot of positive validation, but any new theory has its limits of application. We can fund some research projects specifically to explore the limitations of the third-order model under extreme conditions—such as ultra-low temperature environments, high radiation environments, or special material systems. This is not an 'attack,' but an 'academic discussion.' However, if we can find the limits of the third-order model's applicability, we can gain more say in the standard-setting process."

"You mean—finding scenarios where the third-order model is not applicable?" Stein asked directly.

"Yes. Every theory has boundaries. Finding those boundaries is valuable academic work in itself. It's just—" Bertoli paused, "...it's even more advantageous if those boundaries happen to fall within our area of ​​strength."

Stein did not respond immediately. His fingers tapped silently a few times on the table.

"The second direction," Bertoli continued, "is the standardization path. Currently, there are no unified international standards for modeling in the MEMS industry. If the three of us, together with industry associations, take the lead in developing a new standard for MEMS thermoelastic coupling modeling—we can ensure that our own technical approaches are incorporated during the standard-setting process, while setting reasonable verification thresholds for the inclusion of third-order models."

“A reasonable verification threshold.” Stein repeated the phrase, a subtle implication in his voice. “You mean to make the standardization process for the third-order model as lengthy as possible.”

"What I mean is ensuring the rigor of the standardization process," Bertoli corrected, "but objectively—yes, this will slow down the rate at which third-order models become the industry standard."

Weber nodded thoughtfully on the other side of the screen.

"The third direction—" Bertoli's voice lowered slightly, "is the supply chain. Vilan's current production line relies on etching equipment from Hongyuan Precision Equipment. While Hongyuan's equipment is leading domestically, a significant portion of its core components come from European and Japanese suppliers. If—and I mean if—these suppliers have to choose between our three order demands and Vilan's order demands…"

He didn't finish speaking.

But the meaning is already very clear.

Stein's fingers stopped tapping.

"Marco," his voice changed—not with anger, but with a heavy tone, "I can understand the first two directions you mentioned. Academic exploration and standard setting are legitimate industry practices. But the third direction—pressure on the supply chain—crosses the line."

"I'm just presenting all the possible options," Bertoli said calmly.

"This option is not on my radar," Stein stated clearly. "Bosch's principle is technological competition, not supply chain strangulation. And—frankly—if such tactics were exposed, the reputational damage to all three of us would far outweigh the actual impact on Vilan."

Weber cleared his throat softly on the other end of the screen. "I agree with Stein. The third direction is too risky and the potential rewards too small. Let's abandon it."

Bertoli didn't insist. "Then let's just discuss the first two directions."

……

Over the next twenty minutes, the three had a more specific discussion on two areas: academic research and standards development.

Bertoli proposed that STMicroelectronics fund a team led by Kosuke Yamamoto of Kyoto University to specifically study the applicability of the third-order model to special material systems (non-silicon-based MEMS). Weber indicated that Infineon could propose the development of new standards within the IEC (International Electrotechnical Commission) MEMS working group. Stein agreed to Bosch's participation in the standards discussion but emphasized that Bosch would not oppose the inclusion of the third-order model—only requiring "full verification."

"How long will it take for full verification?" Bertoli asked.

"The standard committee's normal process takes two to three years," Weber said.

"Two to three years." Bertoli repeated the number, a slight smile playing on his lips. Two to three years would be enough time for them to fully digest and absorb the third-order model and establish their own technological barriers.

But Stein noticed Bertoli's expression.

"Marco, I need to clarify something." His voice became more serious. "Bosch's attitude towards third-order models is not confrontation, but cooperation. We are already assessing the feasibility of establishing technical cooperation with Vilan through academic channels. If their paper is formally published, we will propose a cooperation framework in the third quarter."

This statement surprised both Bertoli and Weber.

"You mean cooperation?" Bertolli's tone held a hint of surprise.

"Cooperation," Stein repeated. "Our €180 million, 400mm investment requires the best modeling tools. If the third-order model is currently the best, then we should use it, not fight against it. As for how to obtain it—licensing, joint development, or some other form—these are all negotiable. But the direction is clear."

The call fell into another brief silence.

Weber broke the silence first. "Stein, your stance is clear. I respect it. Infineon's final strategy still needs internal discussion, but I personally prefer to remain proactive in standards setting and open in terms of technology."

"That's a reasonable position," Stein said.

Bertoli remained silent for an even longer period.

"I understand your point of view," he finally spoke, his voice regaining its previous calm. "Italy and France will continue with our own evaluation process. We will remain in communication regarding the standards development."

He did not mention cooperation.

The call ended in a polite but subtle atmosphere.

……

After hanging up the video call, Stein sat in his office for a long time.

He knew Bertolli hadn't given up on the third direction. STMicroelectronics had a deeper supply chain presence in the Asia-Pacific region than Bosch—they had established two-decade-long partnerships with several key Japanese component suppliers. If Bertolli truly decided to exert pressure through the supply chain, Bosch would find it difficult to stop him.

But Stein also knew that such methods were becoming increasingly ineffective in today's business environment. Information spread too quickly, and compliance reviews were too stringent. Any exposed manipulation of the supply chain would trigger regulatory scrutiny and a media storm.

More importantly, Stein had a deep-seated conviction: the third-order model won not because of how strong Wei Lan's business operations were, but because the theory itself was correct.

You can slow down the spread of a correct theory, but you can't stop it.

Therefore, Bosch chose to cooperate.

This wasn't because Bosch was kind or enlightened. It was because Stein had done some calculations—

The cost of obtaining a third-order model license through collaboration with Vilan is far lower than the cost of developing an alternative from scratch. It is also far lower than the losses incurred by missing the 400mm production line optimization window.

Business decisions are never about right or wrong, but about costs.

……

April 20th.

Shanghai, Vilan Microelectronics.

A formal notification letter from Star Packaging Technology was placed on Lin Wei's desk.

"This is to inform you that, in accordance with Article 12 of the Alliance Cooperation Agreement signed by both parties on [Date], our company has decided not to renew the agreement, and our alliance membership will automatically terminate upon its expiration on March 31, 2021."

Lin Wei finished reading the notice and placed it on the corner of the table.

She knew this letter would come. StarChip had verbally informed them of their intention to withdraw in mid-March. The reason was simple—their main customer was STMicroelectronics, accounting for over 40% of their revenue. Staying in Vilan's alliance might affect their relationship with STMicroelectronics.

The number of companies has decreased from 29 to 28.

Is the impact significant?

Not significantly. Xingchen Packaging is not a core member. Their packaging services can be partially replaced by Vilan's self-built packaging line, which will begin production in May.

But this is a signal.

A signal about "taking sides".

If STMicroelectronics does start exerting pressure through supply chain relationships—not direct, explicit pressure, but the kind of subtle, unprovable "hints"—then members of the alliance with deep business relationships with Bosch, STMicroelectronics, and Infineon will face similar choices.

Lin Wei picked up the list of alliance members and examined it line by line.

Of the 28 companies, six have direct or indirect supply relationships with Bosch. Four cooperate with STMicroelectronics. Three do business with Infineon. After removing overlaps, approximately nine to ten companies may face pressure to "choose sides."

This includes Mr. Wang, the general manager of Jingce Microelectronics.

But Mr. Wang stated last month that he had abandoned the idea of ​​"betting on two fronts" and chose to remain in the league. At least for now.

Lin Wei put down the list.

She knew this game wouldn't end quickly. Bertoli of Italy was a patient and shrewd opponent. He wouldn't use obviously overstepping tactics, but he would continuously apply pressure in the gray areas.

But Bosch's Stein—Lin Wei's intuition told her that Stein and Bertoli had different attitudes. At the IEEE MEMS session, the questions Stein asked didn't seem like he was looking for weaknesses to attack, but rather like he was assessing the possibility of cooperation.

But intuition cannot be used as evidence.

Lin Wei turned on her computer and began writing a document. The title was:

"Alliance Risk Assessment and Response Plan (Q2 2021)"

She listed three priorities:

First, accelerate the commissioning of self-built packaging lines to reduce reliance on external packaging services.

II. Communication with core members – Confirm the cooperation intentions and potential sources of pressure of each of the eight core members.

III. Alternative Supply Chain Solutions – Prepare alternative suppliers for key links that may be under pressure.

After finishing writing, she glanced at the time. It was 3:30 PM.

She has one hour free.

She opened another file—the first draft of a patent application that Su Chen had sent yesterday. The first one was about a method for optimizing MEMS thermoelastic coupling process parameters based on a third-order model.

A 34-page technical document.

Lin Wei began to review the pages one by one.

……

April 22nd.

Beijing.

Su Chen received two emails.

The first email came from Vogt. It was a confirmation of a subsequent thematic conference—scheduled for mid-June, lasting half a day, and to be held both online and offline. Seven research groups have confirmed they will present their third-order model reproduction results. Vogt invited Su Chen and Zhou Zhiyuan to each give a 30-minute keynote speech.

Su Chen forwarded the email to Zhou Zhiyuan, adding the line: "June. The timing matches."

Zhou Zhiyuan replied with two words: "Prepare."

The second email came from NM's editing system.

Su Chen's heart didn't race when he opened it—he had trained himself to remain physiologically calm when facing uncertainty.

The email subject is:

"NM-2020-12-25-00847: Status Update"

The content consists of only one line:

"Your revised manuscript has been sent to reviewers for re-evaluation."

The revision has been sent for the second round of peer review.

This is at least a week faster than usual. Typically, NM editors need one to two weeks to confirm the completeness of a revision before sending it for review. This time it only took two weeks.

Su Chen didn't overanalyze it. But he noticed a detail—the email was sent at 2:00 AM London time.

This means that the editor processed this manuscript late at night.

It could just be a work habit. Or perhaps—this paper was placed in a priority queue.

Su Chen closed the email.

He drew a timeline in his notebook:

April 22: Revision submitted for review.

Expected four to six weeks: second round of peer review.

Late May to early June: Results will be available.

If accepted: Officially published in June.

Mid-June: Vogt Thematic Conference.

End of June: Tian Shu chip solution three deadline.

All the clues converged in June.

He closed the notebook.

Outside the window, Beijing in late April had fully entered spring. The leaves of the London plane trees shimmered with a vibrant green luster under the sunlight, and the air was filled with the warm scent of grass.

Su Chen stood up and walked to the window.

He stood there for about thirty seconds.

Then he turned back to his desk and opened the draft of the third chapter of the second paper.

Fingers landed on the keyboard.

The rhythm is steady.

as always.