Cheng Yuan stared at the report on the screen for a full ten minutes, the third time.
<Internal alignment regarding collaboration with Source Code Technology (draft)>.
His private office was at the very back of the strategic investment department, with a glass partition offering a view of the more than twenty workstations outside. The Dingsheng Haicheng headquarters, on the 33rd floor, had the strategic investment department occupying a third of the entire floor. To emphasize its commitment to technological innovation, a previous VP renamed all the meeting rooms on this floor after scientists— Turing Hall, Von Neumann Room, Shannon Hall —and they haven't changed them since. Each floor's standard coffee corner features three fully automatic Coffee Machine, next to which is a notice from the administration department: the coffee beans have been changed to Geisha blend; feedback is welcome.
That's how big companies are; even getting a cup of coffee requires a feedback process.
Cheng Yuan turned his attention back to the report. He revised the last paragraph twice, changing "extremely low feasibility" to "feasibility needs to be verified," and "deep involvement is not recommended at this time" to "it is recommended to first make contact to understand the specific demands."
The wording should be objective; this is a matter of professional ethics.
He leaned back in his chair, took off his metal-framed glasses, and wiped them. At forty-three, he'd been in strategic investment for almost twenty years; he'd seen all sorts of startups. But with Source Code Technology, he did have some personal feelings.
Last month, the internal All Things Alive had already failed, and Boss Zheng asked him to explore the possibility of a complete acquisition.
Cheng Yuan prepared a complete valuation plan, transaction structure, and even an exit mechanism, and arranged to meet Han Luyi at the Source Code office. However, Han Luyi look at the valuation plan, and simply said, "Cooperation is negotiable, but acquisition is out of the question."
It was straightforward and efficient, without offering any leeway.
How old is he? Cheng Yuan. They accidentally got caught in the right trend and then lost their sense of their own limitations. In the end, not many of them could even keep their companies alive for more than three years.
Cheng Yuan smiled professionally and said he understood, adding that he would report back to Boss Zheng that "the other party is not interested at the moment."
Six words were enough to suppress all my discomfort.
Rationally, he also understands that it's only natural for the founders of YiYi not to sell the company.
But as the VP of Strategic Investment at Dingsheng, he hadn't personally gone out to negotiate business in a long time, let alone been treated with this kind of attitude.
However, Boss Zheng didn't press further, only saying, "Keep an eye on it."
Stay tuned.
The "stay tuned" that the CEO of Prosperity Group said is not the same as the "let's have dinner together sometime" that ordinary people say.
Cheng Yuan dared not neglect anything that Boss Zheng took seriously.
Moreover, Chairman Lü recently gave an internal speech stating that the company's resources should be tilted towards AI.
To what extent is this biased? From what level does this focus on Source Code Technology originate?
At the Dingsheng, you can guess, but there's nowhere to confirm it. If you guess correctly, that's your skill.
So ever since then, he has been tracking the public updates of Source Code Technology: Open Things's monthly active user data, the progress of the Qianfan White Label cooperation, and even the number of stars on GitHub for BugKiller's open source project.
A Series A company with a post-investment valuation of 2 billion, with fewer than 50 people in total. To be honest, Dingsheng's model team alone has more than 200 people, and the project team temporarily drawn from All Things Alive had more than 100 people. A company the size of Source Code Technology would normally struggle to even be included in the weekly reports of strategic investors, Cheng Yuan himself would not even have the chance to hear about it.
But Han Luyi is different.
Last month they rejected the acquisition offer so decisively, but this month they've proactively contacted me to discuss cooperation. The messages they sent made Cheng Yuan laugh in exasperation.
He couldn't figure out what Han Luyi was up to, but one thing was clear: the other party wasn't there to surrender.
Such people either have hidden cards up their sleeves, or they are simply arrogant and ignorant.
Cheng Yuan read the report again, confirmed that the format was correct, and submitted it through internal channels.
The report was short, deliberately kept to one page or less. Boss Zheng is extremely busy; he most likely wouldn't read a report longer than two pages.
<Internal alignment regarding cooperation with Source Code Technology (draft)>
I. Background
Source Code Technology CEO Han Luyi recently contacted the Strategic Investment Department, proposing a collaboration in the area of Large Model. After initial communication, the meeting is tentatively scheduled for the afternoon of December 15th (next Tuesday).
II. Brief Description of the Target
Founded in July 2026, Source Code Technology completed its Series A funding round in October, led by Hongyuan Capital, valuing the company at approximately 2 billion RMB post-investment. The company currently has a team of about 50 people. Its products include BugKiller (code inspection SaaS), Open Things (AI application generation platform), and a white-label API distribution business with Qianfan Technology. Open Things a monthly active user growth of approximately 200% in the past three months, and its penetration rate among non-technical users is among the highest in the industry.
III. Preliminary Assessment of Cooperation Intent
Han Luyi mentioned "cooperation in the direction of Large Model," but specific demands are not yet clear (Computing Power /technology/funding, etc., need to be discussed and confirmed in person). IV. Strategic Investment Assessment
Positive aspects: Open Things has established a relatively clear product barrier at the application layer; user behavior closed-loop data has independent value; and it has a first-mover advantage in the market. Risks: The team size is seriously mismatched with the investment in Large Model R&D; there is no experience in developing base models; and the existing funding chain is insufficient to support the Model Training.
V. Recommendations
It is recommended to first reach out to understand the specific needs and assess the possibility of cooperation or strategic investment.
After submitting the report, Cheng Yuan the restaurant for lunch. He estimated that the approval process for this report wouldn't be completed until the end of tomorrow at the earliest.
Unexpectedly, my phone rang while I was sitting by the window drinking coffee after finishing my meal.
Everyone in the strategic investment department knows that Cheng Yuan always keeps his phone on sound and doesn't use vibrate mode so he doesn't miss calls, but not many people know his number, and he only uses Whitelist.
"Good afternoon, Mr. Cheng. Boss Zheng has reviewed the alignment report you just submitted and would like you to come to the 49th floor at 3 PM." It was Boss Zheng's secretary. Cheng Yuan glanced at the time: 1:47 PM. The report had been submitted less than two hours ago, and it was still lunchtime. "Okay, I'll be there on time."
After hanging up the phone, Cheng Yuan was first surprised, then he understood: Boss Zheng's attention to the Source Code issue was even higher than he had anticipated. It was just a draft for strategic investment alignment; a routine email approval would have sufficed. The fact that he personally met with him showed that he truly cared.
He quickly returned to his office to print two reports, and then rummaged through his drawer to find the supplementary materials he had prepared for Source Code Technology last month: product screenshots, competitor analysis, and publicly available user growth data. He put them all into a folder to prevent any omissions.
As the time approached, Cheng Yuan straightened his tie and set off ten minutes early.
The 49th floor, the floor below the top floor of Dingsheng Tower.
Zheng Xiaobo's office is located in a corner, with floor-to-ceiling windows facing the Huangpu River.
Pushing open the door, the first thing you see is a desk. The office is unbelievably simple.
Cheng Yuan had been to this Cheng Yuan least twenty times, and each time he felt the same way: unlike Dingsheng, it lacked that money-driven atmosphere. "Sit down."
Zheng Xiaobo sat behind his desk, on which lay a printed document—the one he had submitted that morning. Cheng Yuan noticed that there were no annotations on it.
"You met Han Luyi last time, what are your thoughts?"
Zheng Xiaobo didn't look at the report; he just asked people directly.
Cheng Yuan was prepared. "Their product is indeed quite impressive. Open Things has differentiated itself at the application layer, especially in the non-technical user market, which they've done very cleverly. Their growth data is also very impressive, placing them among the top performers in this year's AI field."
Cheng Yuan didn't mention the competition between Open Things and All Things Alive. Although it was a project led by Wang Zhiyuan, it was still under the name of Dingsheng. It was not a good thing to say it out loud, so no one dared to discuss it openly in Dingsheng. However, he guessed that Boss Zheng used this opportunity to defeat Wang Zhiyuan.
Cheng Yuan paused for Cheng Yuan, then added, "However, this person is not easy to deal with. He rejected the acquisition proposal last time without even considering it."
Zheng Xiaobo picked up his water glass, took a sip, and made no comment.
Cheng Yuan understood Cheng Yuan. Boss Zheng didn't care about his pride.
"Cheng Yuan, I have a question for you." Zheng Xiaobo tapped the table with his finger. "What exactly does Dingsheng want?"
The question wasn't directed at Cheng Yuan, Cheng Yuan could tell that.
Zheng Xiaobo stood up, walked to the French windows, and spoke with his back to Cheng Yuan. The Huangpu River looked grayish under the winter sun.
"The All Things Alive's application has crashed. Our approach of building our own application layer didn't work."
Cheng Yuan didn't interrupt. He couldn't bring up the matter of All Things Alive, but since Boss Zheng brought it up, it meant that he was setting the tone.
"What Han Luyi accumulated at the application layer—product methodology, user mindset, data loop—that's what's truly valuable," Zheng Xiaobo turned around, his gaze calm. "It's best if we can directly utilize his expertise."
Cheng Yuan nodded. Acquisition, or deep control. This was also the direction he hinted at in his report.
"Boss Zheng, there's one more detail." Cheng Yuan leaned forward. "When Han Luyi contacted us, he mentioned the Large Model direction." "Large Model direction?"
"Yes, it's not just about selling products or getting investment; he might want to Large Model."
Zheng Xiaobo had his back to Cheng Yuan, and a smile had already appeared on his face.
That's a good joke.
A company of fifty people with Series A funding, and they're already Large Model? Do you know how much Dingsheng spent on the Ganyuan? Do you know how much budget Dingsheng has for their current Kun Yuan Large Model?
He didn't turn around, his tone flat. "Large Model? Does he even know what level of investment that entails?"
Cheng Yuan doesn't need to answer.
Overestimating one's abilities.
No one said it aloud, but it was the unspoken assessment between the two of them.
Zheng Xiaobo walked back to his desk, his smile gone. This wasn't something he should let Cheng Yuan see.
"Dr. Liu's project has yielded some preliminary results. It just passed its internal review last week, and several key indicators are quite good. I haven't looked at the specific figures in detail, but the feedback from the model team has been very positive."
Dr. Liu Dahai, a top-notch technical expert poached from OpenAI by Dingsheng two years ago, is now the Chief Scientist in the model group. Cheng Yuan know why Zheng Xiaobo suddenly brought this up.
Zheng Xiaobo nodded at him. "When he comes next Tuesday, have Dr. Liu participate as well, and give a brief technical demonstration."
Cheng Yuan immediately understood. This was a display of Dingsheng technical prowess, letting Han Luyi see firsthand where the gap lay.
"He's a good person, and he has ideas for the product." Zheng Xiaobo picked up his water glass and concluded, "But if he's going to make an app, he should just focus on making a good app. Let him see what happens, and give up on this idea as soon as possible."
The final sentence sets the tone:
"If he changes his mind after the meeting, we can discuss the rest properly, within your authority. If it's not enough, come to me."
Cheng Yuan stood up, accepted the order, and left.
After talking with Boss Zheng, he had a direction: cooperation on Large Model was out of the question, but company acquisition was still negotiable, and the price could be doubled. This was within his authority, and negotiating within this range would be his achievement.
As the elevator doors closed, Cheng Yuan took out his phone and sent a message to his secretary:
"Could you connect me with Dr. Liu Dahai? There's an external meeting next Tuesday afternoon. Boss Zheng would like Dr. Liu to give a brief technical demonstration of the base model. Please confirm the time with his assistant."
The secretary replied instantly, "Yes, Mr. Cheng, I'll contact him right away."
A few minutes later, before the elevator even reached the first floor, the secretary sent another message: "Dr. Liu confirmed it, asking what direction the demonstration should take?" "It's about the base model. The other party is a startup company making AI applications, and they want to discuss cooperation on Large Model."
"Received, I'll send out the meeting invitation."
Cheng Yuan put his phone back in his suit pocket and watched the elevator numbers jump to thirty-three.
Next Tuesday, things are going to get interesting.