Tech Giants: Starting from Bankrupt Companies

143: Chapter 143 A Symphony of Technology: Star AIHangzhou Future Sci-Tech City, December 17th, 9:00 AM.

In Hangzhou Xixigu, in front of a brand-new twelve-story glass-curtain building, the ground was covered with a red carpet and surrounded by fresh flowers.

The golden plaque of the ”Xingchen Artificial Intelligence Research Institute” was covered by a red silk cloth, waiting to be unveiled.

More than five hundred people had gathered in the square downstairs—government officials, university professors, corporate representatives, media reporters, and over a hundred fresh graduates waiting in line for interviews.

Lin Yusheng stood on the temporarily erected rostrum. At 38, she wore a dark blue suit, her short hair was neat, and her makeup was exquisite.

However, those who knew her well knew that this former Chief Scientist at Microsoft Research Asia hated formalism the most.

”Dean Lin, five minutes left,” the assistant reminded her.

Lin Yusheng nodded, pulled a ballpoint pen from her pocket, and began to disassemble it unconsciously—cap, refill, spring, barrel... She disassembled it into parts within ten seconds and reassembled it in five.

This was her habit. At Microsoft, before every important technical review meeting, she would disassemble and reassemble a pen like this to calm her nerves.

But today, she wasn’t doing it to calm her nerves; she was thinking about what she was about to say.

Su Chen had made the positioning of the Xingchen Artificial Intelligence Research Institute very clear: ”We do not do paper AI, we do not do demo AI; we want to do industrial AI that can enter factories, go down to production lines, and be integrated into chips.”

This positioning suited her perfectly. During her ten years at Microsoft Research Asia, she had published 87 top-tier conference papers, won three ”Best Paper Awards,” and mentored three IEEE Fellows.

But in the dead of night, she would always ask herself: Did these studies really change the world?

Most did not. They stayed in papers, in demo videos, and in reports by tech media.

But Xingchen provided her with a completely different platform: the real factories, real production lines, and real data of the xinghuo system.

There was no need for showing off skills here, only for solving problems.

”Time’s up,” the assistant said.

Lin Yusheng put away the pen and stepped onto the rostrum.

Applause rang out. She bowed slightly, without taking any speech notes.

”Good morning, everyone. I am Lin Yusheng, Dean of the Xingchen Artificial Intelligence Research Institute.” Her voice carried across the square through the speakers.

”Before the unveiling, I want to ask a question: What should artificial intelligence really be?”

The audience was quiet.

”Is it winning against humans in Go? Is it writing poetry or painting? Is it chatting with people?” Lin Yusheng shook her head.

”These are certainly important, but they are not everything. In my opinion, AI should be an optimizer of the physical world, a solver of engineering problems, and a tool to make machines smarter, production more efficient, and life better.”

The big screen behind her lit up, showing two project names:

”Yu” Project: AI-Empowered chip design

”Luban” Project: AI-Empowered Intelligent Manufacturing

”These are the two core topics of the first phase of the Xingchen AI Research Institute,” Lin Yusheng introduced.

”The goal of the ’Yu’ Project is to shorten the design cycle of 7nm chips from 18 months to 8 months and increase the yield prediction accuracy to 99.7%. How will we do it? By using AI to learn tens of thousands of rules of chip design, automatically optimizing layout and routing, and predicting manufacturing defects.”

A low gasp of astonishment rippled through the audience. chip design is an intelligence-intensive industry; a senior engineer needs ten years of experience to design independently. Now, AI wants to intervene?

”The goal of the ’Luban’ Project is to establish a ’digital twin’ of the automotive welding production line and perform real-time tuning, increasing the production rhythm of the Red Flag H-series by 25%,” Lin Yusheng continued.

”This means that the same factory can produce 100,000 more cars per year.”

She paused, looking at the audience: ”I know some people are skeptical: Can AI really do these things? My answer is: How will you know if you don’t try?”

”Xingchen has provided us with the most precious thing—real scenarios, real data, and real needs. This is a hundred times more valuable than any simulated environment in a laboratory.”

”Therefore, the recruitment standards for the Xingchen AI Research Institute are very simple,” Lin Yusheng pulled up the recruitment page.

”First, do not be afraid of difficult problems. Second, do not be afraid of going into the factory. Third, do not be afraid of 996—of course, we will provide sufficient compensation.”

The compensation plan was displayed on the big screen:

Doctoral fresh graduates: Starting salary 600,000/year + project dividends

Master’s fresh graduates: Starting salary 400,000/year + project dividends

Part-time professors: Retain university tenure, hourly wage starting at 3,000 yuan

This figure caused an uproar in the audience. An Alibaba P7-level engineer’s annual salary is roughly 800,000-1,200,000, but that requires over five years of experience.

Xingchen offering 600,000 to fresh doctoral graduates is almost the top salary in the industry.

More importantly, there were ”project dividends”—if the ”Yu” or ”Luban” projects were successful, the generated benefits would be distributed to team members proportionally.

Taking chip design as an example, shortening the design cycle by 10 months meant hundreds of millions or even billions in market opportunities.

”Now, let’s unveil it.” Lin Yusheng stopped talking, walked to the plaque, and pulled away the red silk with the Hangzhou city leaders.

The seven large characters ”Xingchen Artificial Intelligence Research Institute” shone in the sunlight.

Thunderous applause. But Lin Yusheng knew the real challenge was just beginning.

As soon as the unveiling ceremony ended, a long queue formed in front of the recruitment booth.

It wasn’t dozens, or hundreds, but over a thousand people—the line stretched from the building entrance to the street corner, and people kept joining.

”Classmate, how long have you been waiting?” a reporter asked randomly.

”I’ve been here since six in the morning,” a boy wearing glasses replied.

”I’m a doctoral student in the Computer Science Department at Zhejiang University, graduating next year. Xingchen’s ’Yu’ project is exactly my research direction—I’ve been using AI for circuit optimization for three years, but I’ve never had the opportunity to access real chip design data.”

”Why not go to Alibaba or Tencent? Their AI labs are also very famous.”

”It’s different,” the boy shook his head. ”The AI at Alibaba and Tencent is mainly in e-commerce and social networking, which is software-level optimization. Xingchen’s AI is a combination of software and hardware, which can truly change the physical world. I study engineering, and I want to do things that can be implemented.”

Such conversations were everywhere in the queue. There were not only fresh graduates in the queue but also senior engineers from BAT, professors from universities, and even a few Chinese scientists who flew back from Silicon Valley.

In the middle of the queue, Professor Wang, the director of a national key laboratory at Shanghai Jiao Tong University, was waiting with his resume.

He was 58, gray-haired, had been researching in the field of AI for Science for five years, had published in Nature sub-journals, and had won the National Science and Technology Progress Award.

But now he wanted to change jobs.

”Professor Wang, have you really thought about this?” his assistant asked in a low voice. ”Jiao Tong University gave you tenure, and you have your own laboratory...”

”What’s the use of a laboratory?” Professor Wang smiled bitterly.

”For five years, my team has published 27 top-tier conference papers and applied for 15 patents. But all these studies have stopped at the simulation stage—because we don’t have real industrial data or a real testing environment.”

He looked at the Xingchen building: ”But they have it here. The chip design data of Xinghuo, the automotive production data of Red Flag, the machine tool operation data of Xingchen Heavy Industry... This data is more valuable than any simulation. With this data, my research can truly be implemented.”

”But the salary...”

”Salary is certainly important, but the opportunity is more important,” Professor Wang’s eyes lit up.

”Having done scientific research all my life, my biggest regret is that my research results could not be turned into products. Now that there is a chance to make up for this regret, I don’t want to miss it.”

He paused: ”Besides, with a part-time consulting fee of 3,000 yuan per hour, the operating funds for my laboratory will be solved. I can bring the whole team here, doing applied research at Xingchen while leading students to do basic research at Jiao Tong University—this is a win-win.”

The assistant stopped persuading him. She knew Professor Wang had made up his mind.

At the front of the queue, interviews had already begun. The first interviewee went in for ten minutes and came out with a face full of excitement: ”Passed! Directly given P8 level, annual salary 800,000 plus dividends!”

This news spread through the queue like wildfire. People became more restless and expectant.

On the twelfth floor of the building, in Lin Yusheng’s office, Li Jianguo, the Vice Dean of the School of Computer Science at Zhejiang University, was sitting on the sofa opposite her.

”Dean Lin, you are trying to hollow out Zhejiang University,” Li Jianguo said half-jokingly.

”Among our graduates next year, 8 out of the top 10 listed Xingchen AI as their first choice. Yesterday, the HR Director of Alibaba even called me to ask if I could persuade the students.”

”Dean Li, this is not poaching; it’s a win-win,” Lin Yusheng poured him tea.

”Xingchen needs talent, and universities also need feedback from the industry. We can establish a joint laboratory; students can intern at Xingchen, professors can work part-time, and research results can be quickly transformed.”

”That’s easy to say, but the salary gap is too big,” Li Jianguo shook his head.

”An annual salary for one of our professors is only 300,000 to 400,000. You offer part-time consultants an hourly wage of 3,000; coming for four days a month exceeds that. Over time, if professors all run off to do part-time jobs, who will focus on teaching?”

Lin Yusheng was silent for a few seconds: ”Dean Li, I understand your concern. But have you ever thought about why universities cannot retain talent? It’s not because companies pay more, but because companies can provide talent with a bigger stage, faster growth, and more direct value realization.”

She stood up, walked to the window, and looked at the kilometer-long line still waiting downstairs: ”These young people have studied hard for twenty years, not to spend their whole lives writing papers, but to change the world. If universities cannot provide them with a path to change the world, they will naturally go to places that can.”

Li Jianguo was speechless. He knew Lin Yusheng was right.

University evaluation systems are oriented toward papers, while corporate evaluation systems are oriented toward output. In today’s rapidly changing era, the latter is clearly more attractive to ambitious young people.

”So, instead of resisting, why not cooperate?” Lin Yusheng turned around. ”Xingchen can sign a strategic agreement with Zhejiang University: we provide data, scenarios, and funding, and Zhejiang University provides talent, basic research, and academic resources. After graduation, students can choose to come to Xingchen or go to other companies, but they will have already been exposed to real industrial problems during their time at school.”

This proposal moved Li Jianguo. School-enterprise cooperation is nothing new, but cooperation like this, where Xingchen is willing to open up core data and real-world scenarios, is rare.

”I need to discuss this with the school leadership,” he said. ”But in principle, I support it.”

”That’s good,” Lin Yusheng smiled. ”Additionally, I have another proposal: Xingchen will establish the ’Xingchen Scholarship,’ with 100 spots annually, covering all stages from undergraduate to doctoral. Recipients must promise to work domestically for at least three years after graduation.”

Li Jianguo’s eyes lit up: ”This is great! Many excellent students go to America as soon as they graduate. If we can provide competitive development opportunities, they will be more willing to stay.”

The two talked for another half hour and reached a preliminary agreement on the cooperation.

After seeing Li Jianguo off, Lin Yusheng returned to the office. The assistant handed her a report: ”Dean Lin, we received 1,873 resumes this morning, including 412 from doctoral candidates, 961 from master’s candidates, and 500 from experienced professionals with industry experience. 300 interviews have been scheduled, 87 have been completed, and the pass rate is... 65%.”

”Too high,” Lin Yusheng frowned. ”Raise the standards. We are not hiring people to do chores; we are hiring people who can solve problems. Keep the pass rate below 40%.”

”But that will slow down the progress...”

”Let it be slow. Better to have fewer high-quality people than many mediocre ones,” Lin Yusheng said firmly. ”The first batch of teams at the Xingchen AI Research Institute must be the elite of the elite. We must produce results within 100 days, and there cannot be any weak links.”

The assistant nodded and noted it down.

Lin Yusheng walked to the whiteboard, where the architecture diagram for the ”Xingchen AI Large Model” was already drawn. This was not a general-purpose large model, but a vertical model focused on the industrial field.

The architecture was divided into three layers: Bottom layer: Data layer. Integrating the design data of Xinghuo Chip, production data of Spark Materials, manufacturing data of Hongqi Automobile, and operational data of Xingchen Heavy Industry... real-time data streams from all subsidiaries.

Middle layer: Learning layer. The AI model learns from this data and discovers cross-domain optimization patterns—for example, how the heat treatment process of materials affects the yield of Xinghuo Chip, and how the vibration patterns of machine tools affect the precision of auto parts.

Top layer: Application layer. Applying the learned patterns back to each subsidiary: optimizing material formulas, optimizing chip designs, and optimizing manufacturing processes to form a positive cycle.

This was a grand vision. If successful, the xinghuo system would no longer be a collection of independent companies, but an organism unified and coordinated by an AI brain.

But the difficulty was also immense. Data formats varied wildly, security levels differed, and real-time requirements were extremely high. More importantly, industrial problems often lacked standard answers, requiring the AI to find optimal solutions under constraints—which was much more complex than playing Go.

”Dean Lin, a video conference with General Manager Su,” the assistant reminded her.

Lin Yusheng connected. Su Chen’s face appeared on the screen, with his office in the background.

”Yusheng, did the unveiling go smoothly?”

”Smoothly. We received nearly two thousand resumes, and professors from Zhejiang University and Shanghai Jiao Tong University came.”

”Very good,” Su Chen nodded. ”But I only care about one question: when can we have the first quantifiable output?”

Lin Yusheng had anticipated this question. Su Chen was a results-oriented person who wouldn’t pay for visions, only for results.

”100 days,” she said. ”Give us 100 days. In 100 days, the ’Yu’ project will provide a set of chip design optimization solutions, with the goal of reducing the power consumption of a certain chip by 10%. The ’Luban’ project will provide a set of automotive welding line optimization solutions, with the goal of increasing production pace by 5%.”

”The target is too low,” Su Chen frowned.

”This is the first phase target,” Lin Yusheng explained. ”We must first prove that the methodology is feasible, and then scale up. If we pursue a 20% or 30% improvement in the first phase, the risk of failure is too high.”

Su Chen thought for a few seconds: ”Alright. But in 100 days, I want to see concrete data, not demo videos.”

”Understood.”

”Additionally, pay attention to talent security,” Su Chen reminded her. ”The Xingchen AI Research Institute is currently in the spotlight, and various forces will want to infiltrate it. Recruitment must undergo strict background checks, and the core team must sign non-compete and confidentiality agreements.”

”We are already doing that.”

”Good. See you in 100 days.”

The video call disconnected. Lin Yusheng leaned back in her chair and let out a long sigh.

100 days. A little over three months. To use AI to solve the two most complex industrial problems—chip design and automotive manufacturing—and to produce quantifiable results.

The pressure was immense.

But she liked this pressure. When she was at Microsoft, her biggest regret was that her research results could not be implemented. Now, the opportunity had come.

She picked up the ballpoint pen and began to disassemble and reassemble it again. The pen turned into parts in her hands and then back into a whole, over and over again.

Just like what she was about to do: break down complex industrial problems into data, use AI to find patterns, and then reassemble them into solutions.

Disassemble, understand, reorganize, optimize.

This is the philosophy of AI, and also the philosophy of industrial progress.

Downstairs, the recruitment line was still extending. Upstairs, a brand-new AI era was quietly setting sail in this wintery Hangzhou.

At 23:00 late at night, the core figures of the xinghuo system in Shenzhen, Hangzhou, and the Imperial Capital had not yet slept.

In Shenzhen, Xu Lihua was still in the operations center staring at sales data: the number of orders had exceeded 150,000 units, and sales had surpassed 800 million. He ordered the production lines to start working in three shifts and simultaneously began preparing logistics plans for overseas shipping.

In Hangzhou, Lin Yusheng was still interviewing the last candidate—a chip architect who had returned from Silicon Valley with twenty years of experience and was willing to take a pay cut to join Xingchen because he ”wanted to participate in something that truly changes the industry.”

In the Imperial Capital, Wu Yi was watching surveillance footage in a secret room. Chen Zhe had just left the laboratory, holding a USB drive—it contained some of the technical documents for the ”Blue Bird Project.” He walked to the parking lot, sat in his car for ten minutes, and then drove to a public parking lot in the city.

He knew there was a dead drop there, the place where Chen Zhe passed on intelligence.

But he did not intend to intercept it. Let the intelligence be passed on, let the fish take the bait.

At the same time, Su Chen stood in front of the floor-to-ceiling window of his office, holding a report that had just been printed.

The cover of the report read: ”xinghuo system 2018-2028 Strategic Plan”

Inside, it detailed the development paths for the five major sectors—materials, chips, batteries, automobiles, and AI—over the next ten years, as well as—most importantly—the construction schedule for a completely autonomous supply chain.

According to the plan, by 2020, the xinghuo system would achieve complete autonomy from raw materials to terminal products, no longer relying on any key foreign technologies.

This was an almost impossible goal to achieve.

But Su Chen knew it had to be done. Because if it wasn’t, today’s victory would only be a flash in the pan. The counterattack from opponents would be more fierce than imagined.

He picked up a pen and wrote on the last page of the report: ”In the war of technology, there is no retreat. Only by forging ahead can we find a way out of a desperate situation.”

Then he signed his name.

Outside the window, the night sky of Shenzhen was dyed dark red by the city lights. This city, this country, was experiencing an unprecedented technological revolution.

And Xinghuo was the ignition point for this revolution.

Su Chen turned off the office lights but did not leave.

He stood in the darkness, recalling the scene two years ago when he first explained the Xinghuo vision to Ren Zhengfei.

Ren Zhengfei asked at the time: ”Su Chen, do you know how difficult this path is?”

He replied: ”I know. But someone has to walk it.”

Ren Zhengfei was silent for a long time, then said: ”Good. Then let’s walk it together.”

Now, they were already on the road. Ahead lay fog, treacherous waters, and unknown challenges.

But behind them was the future of the entire Huaxia high-tech industry that was being lit up.

Su Chen took out his phone and sent the last message: ”All units, proceed according to plan. Tomorrow, a new day.”

After sending the message, he looked to the east.

In a few hours, the sky would brighten.

A new day, a new battle, was about to begin.