Monday, 9:30 AM, Pengcheng, Nanshan District.
Lu Mingzhou sat in the car, looking out the window at a gray-white research and development building.
The building wasn't tall, and the glass curtain wall was so dirty it couldn't be wiped clean. There was a plastic nameplate hanging at the entrance.
Sili-Lumin Semiconductor.
The English name, Sili-Lumin Semiconductor, is also written next to it.
Looking at this building alone, no one would associate it with the term "new force in domestically produced AI Chip".
The whole scene looked like a dilapidated factory.
Lu Mingzhou got out of the car, followed by his assistant Wu Tong.
Wu Tong was a CFO office assistant hired by Lu Mingzhou after the Chinese New Year. He had a master's degree in finance and had previously interned in the M&A group of a securities firm. His resume was quite impressive, but this was his first time actually sitting at a negotiation table for an acquisition.
Wu Tong stepped forward and whispered, "President Lu, Ximing's management team will also be present today."
Lu Mingzhou turned around and glanced at him.
So what if it is?
Today's main negotiating partners are several existing investors who collectively hold 53.6% of the shares. If they are willing to transfer their stake, Source Intelligence can directly become the controlling shareholder.
Yes, Source Intelligence has been established for so long, Han Luyi not assembled a new team for Source Intelligence.
Lu Mingzhou sensed that he would likely plan to merge the two companies in the future.
At that time, Lu Mingzhou's shares in Source Code could be exchanged for shares in Source Intelligence.
It's important to understand that while the Source Code's valuation is quite good, Source Intelligence is currently riding a wave of opportunity, ready to take off at any moment.
Therefore, Lu Mingzhou was determined to make this acquisition a clean and successful one.
He had been researching Sili-Lumin Semiconductor for several days before coming here, and had also asked Jiang Songran and Zhang Biao for details.
Their core product, the L100 graphics card, has high -Hardware and strong performance, but the software is terrible, resulting in poor sales and tight cash flow.
The only large order we received recently came from Source Intelligence Technology.
This briefing alone reveals that this company is a disaster for most investors.
However, this presented a good opportunity for Source Intelligence Technology to take advantage of a loophole.
What Source Intelligence lacks most right now is Computing Power, while Ximing has patents, well-designed products, a well-established supply chain, and even a backlog of inventory.
Most importantly, they have a group of engineers who have actually worked on AI Chip Hardware.
As for their poor software ecosystem and almost non-existent training and deployment support, which buyers can only use for inference of small models with extremely low utilization rates, these are not disadvantages for Source Intelligence.
Since Jiang Songran and his team had already completed the Operator Adaptation for L100, Source Intelligence had no intention of using their software stack at all.
In fact, the problem faced by Ximing is not just its own problem, but a common problem for all domestic Chip manufacturers, and even a common problem for all non- Nvidia manufacturers.
It has Hardware, but no software, and most importantly, no ecosystem.
Without an ecosystem, no one will use it; if no one uses it, no one will buy it, and there will be no ecosystem at all.
However, those large manufacturers have other business lines to subsidize Hardware, or some companies have received more money from investors, so they can still afford to burn through the money.
But Ximing can no longer be burned.
Lu Mingzhou didn't come here today to buy a card.
He wants to buy the entire company outright.
To be precise, it's about buying equity, or more precisely, buying existing shares held by early investors.
Ximing's early investors have lost patience. In the last round of financing, Ximing was valued at 2 billion, but now, let alone 2 billion, even 1 billion, no one is willing to take over.
These institutions are genuinely willing to cut their losses by half, which is perfect timing, as they want to leave and Source Intelligence wants to join, making it a perfect match.
As for the founding team, according to Han Luyi's idea, it would be best if they could stay. However, the current CEO may be very good at Hardware, but his ability to develop software and run a business is really lacking. Otherwise, he wouldn't have been able to manage Ximing to this point.
Ideally, we should let him stay and continue to lead Hardware development.
Lu Mingzhou and Wu Tong had already entered the building, where a receptionist was sitting and playing on her phone.
Wu Tong coughed lightly, and the receptionist finally reacted, giving him a sidelong glance before leading the two to the conference room on the second floor.
The conference room was already quite full, packed with investors from several institutions.
On the other side, Ximing's founding team was also present, with CEO Xue Zhaoheng, who has a technical background, sitting next to his CFO and legal counsel.
The founding team and investors sat separately on opposite sides, clearly indicating that their relationship had soured.
Upon seeing Lu Mingzhou enter, the investors all stood up; on Ximing's side, Xue Zhaoheng also stood up, but his movements were noticeably slower.
"Mr. Lu, it's an honor to meet you." The first investor extended his hand. "Tangyuan Large Model is really making waves lately. I'm Wang Xian, a partner at Guanlan Capital."
"Mr. Wang, you're too kind." Lu Mingzhou shook hands, and the other investors behind him also came over to greet him and shake hands.
After exchanging pleasantries with the investors, Lu Mingzhou went over to the founding team.
Although Xue Zhaoheng and the others greeted them politely, their tone was not very enthusiastic.
After exchanging pleasantries, everyone took their seats.
Wang Xian was clearly the main speaker for the investors today.
He first had someone turn on the projector, and a beautifully prepared document appeared on the screen.
<Explanation of Sili-Lumin Semiconductor Business Development and Equity Transfer>.
There's also a subtitle. "Domestic AI Computing Power Base: Welcoming the Era of Large Model."
Lu Mingzhou saw this and remained outwardly calm, but inwardly he was secretly laughing.
Wang Xian began his introduction.
"Sili-Lumin Semiconductor is one of the few startups in China with the capability to mass-produce high-end AI Chip. The L100 has already entered the mass delivery stage. Although the market environment has been unfavorable over the past year, the company's fundamentals are sound."
After listening, Lu Mingzhou didn't say anything, but silently refuted it in his heart.
It's not that the market environment was bad in the past year; it's that the environment has been bad for the past few years.
"As of the end of last month, the company had signed orders worth 190 million yuan, had 460 million yuan in pending orders, and had a pool of over 30 potential clients. Compared to last year, sales revenue in the first quarter of this year increased by more than 150% year-on-year."
It's only April this year, and the number of signed orders hasn't been explicitly stated. Of all the orders placed since the company's inception, 50 million were placed with Source Intelligence. Of the remaining 140 million, it's unknown how many will actually be delivered or how many will actually receive payment. As for potential orders, there's not even a hint of them.
"With the increasing demand for domestically produced Computing Power from Large Model companies, Ximing will enter a period of performance release this year, and shipments are expected to see a significant leap."
Regardless of whether the industry trend assessment is correct, this is a typical case of mixing industry trends with manufacturer trends. If the increased demand is absorbed by the leading manufacturers, Ximing's situation will likely become even more difficult.
Wang Xian spoke for more than ten minutes, covering topics from domestic substitution to the wave of Large Model, policy opportunities, the performance advantages of L100, and finally, valuation.
Lu Mingzhou listened quietly, his expression remaining unchanged.
"In the last round, Ximing valued at 2 billion. This time, considering the market environment, we are willing to make concessions. Our several funds collectively hold 53.6%, which can be transferred at a valuation of 1.8 billion."
Wang Xian looked at Lu Mingzhou as he finished speaking.
"Mr. Lu, an 1.8 billion valuation is already a very sincere price."
1
Upon hearing this, Lu Mingzhou smiled.
We're all old foxes, why are you trying to play these ghost stories with me?
I've facilitated at least a hundred corporate financing and M&A deals, and the materials and packaging techniques I've worked on are far more sophisticated than this.
This kind of financial manipulation might fool laymen, but don't try to show it off in front of me.
Lu Mingzhou took out the materials at hand and started flipping through them.
"Mr. Wang, I have a question. You just mentioned that you've signed orders worth 190 million yuan. My investigation shows that the total shipment volume of L100 last year was just over 900 units, and the revenue mentioned on the slides was 127 million yuan. Could you please confirm this?"
He asked Wang Xian the question, but his eyes were fixed on Ximing's CFO.
The CFO glanced at Xue Zhaoheng, then at Wang Xian, before speaking, "Yes, it's 127 million yuan."
"How much of that was recovered?"
The CFO remained silent, so Wang Xian interjected, "The collection rate has reached over 80%. The Chip industry has relatively long payment terms, especially since our clients are mainly state-owned enterprises, whose internal processes tend to be slower."
Lu Mingzhou didn't pick up on that topic, but instead pressed on, "The confirmed revenue for the first quarter of this year was 53 million?"
This time, Ximing's CFO nodded.
Last year, the total was less than 130 million, but this year, it reached more than 50 million in just one quarter, which seems like a decent increase. However, don't forget that Source Intelligence made a single purchase of 50 million, while Ximing only sold 30 graphics cards in three months.
What kind of year-on-year growth rate is this, 150%?
"Mr. Wang, I think we should change our pricing logic," Lu Mingzhou said, turning to Wang Xian. "The previous valuation was meaningless. We can give it a suitable valuation based on the company's existing assets and cash equivalents."
Wang Xian's smile froze on his face.
This CEO Lu doesn't play by the rules.
"Mr. Lu, like Source Intelligence, Ximing is a technology company, and the asset pricing method is not applicable to our industry."
Lu Mingzhou leaned his upper body closer to the table.
"Mr. Wang, I believe you wouldn't joke with your reputation; the figures you provided are all true," Lu Mingzhou said. "We'll use these figures as the basis for our calculations."
Then he asked Wang Xian to rewind to the previous slide about sales data.
Lu Mingzhou could tell that these figures had been manipulated; the timing of revenue recognition might have been adjusted, or bad debts might not have been written off. But no matter how they disguised them, they made a fatal mistake.
Or rather, they only packaged it for "other potential buyers," but didn't tailor it to Source Intelligence.
"Our revenue in the first quarter of this year was 53 million, of which Source Intelligence purchased 500 L100 cards, amounting to 50 million."
Upon hearing this, Wang Xian's expression immediately turned sour.
Indeed, others may not know this number, but Source Intelligence certainly does.
"In other words, excluding orders from its own Source Intelligence, Ximing's external market sales revenue in the first quarter of this year was only about three million."
"Three months, three million!"
What does this mean for an AI Chip company that was valued at two billion in its last funding round?
Wang Xian reluctantly replied, "President Lu, you can't look at it that way. Source Intelligence is also a customer, and one of the most representative customers in the Large Model industry. Source Intelligence's purchase itself proves the value of the L100 product."
Lu Mingzhou ignored his desperate struggle. "I come from Auditing background. When I was Auditing, if a company's largest customer accounted for more than 50% of its total revenue, it was considered a major deficiency in operational independence. Right now, Source Intelligence's orders account for more than 50%, don't they?"
"Without this order from Source Intelligence, or even with this order from Source Intelligence, how long will Ximing's Runway last?"
The so-called runway is to calculate how long a company can survive based on its current cash reserves and monthly cash burn rate.
Normally, a company's runway needs to be maintained for more than twelve months to be considered healthy, because this means that the company still has room for trial and error and to change direction. Six months is the warning line, and management needs to raise funds to survive. Three months is already a slow death for the company.
Lu Mingzhou didn't have a specific figure for Ximing's cash reserves; that would only be available during the due diligence phase. However, based on his own model, he estimated it to be less than a month's worth.
This company is on the verge of collapse.
Ximing's CEO, Xue Zhaoheng, remained expressionless throughout, seemingly preoccupied with something else and not fully focused on the negotiations concerning the company's ownership.
Wang Xian remained silent for a long while before finally speaking with difficulty, "President Lu, the total revenue of 127 million last year, that can't be fake, can it?"
"Of course, I believe this revenue is real." Lu Mingzhou leaned back slightly, resting against the back of his chair, and calmly replied, "Of this revenue, how much is from government pilot procurement? How much is from procurement by universities and research institutes? How much is from repeat purchases by corporate clients?"
"Mr. Wang, it's still uncertain whether the government subsidy policy for Computing Power procurement in Pengcheng last year will continue this year."
Wu Tong glanced at Lu Mingzhou sideways, surprised that he would approach the matter from this angle.
How much of last year's revenue was driven by female subscribers? Will it be the same this year? We can already get a glimpse of the situation from the first quarter's follow-up sales.
"President Lu, your way of negotiating seems very insincere." Wang Xian said this while letting out a few laughs.
Lu Mingzhou said, "President Wang, this equity transaction isn't absolutely necessary for us. To be honest, what we're interested in is Ximing's inventory. If we can't reach an agreement, I can use that money to purchase graphics cards from other companies."
That's actually quite blunt – Ximing has practically no valuation left besides its excess inventory, so it might as well be sold off as a package deal.
"Speaking of inventory," Wang Xian tried to make a last-ditch effort, "Ximing currently has at least eight thousand L100 units in stock, which is worth eight hundred million."
"If we can sell them," Lu Mingzhou said, pretending to be a woman.
If you could sell it, would you still be sitting here talking to me?
After exchanging glances with investors from several other institutions, Wang Xian looked at Lu Mingzhou and asked, "Mr. Lu, how much are you prepared to invest?"
"I can acquire your shares based on a valuation of 800 million."
Lu Mingzhou submitted his offer.
This is actually much cheaper than his Xin Ya.
But buying a company is like buying vegetables at a market; you ask for the sky's the limit, then haggle. You can't just reveal your bottom line right away.
"Mr. Lu, you should always leave some room for maneuver in business."
"I've already left it." Lu Mingzhou said, "Has Ximing finished paying off the debts owed to its suppliers? In a few months, if no other money comes in, and they really go into bankruptcy liquidation, will they still have 800 million left?"
Wang Xian stared at Lu Mingzhou for a while, then exchanged glances with the other investors before standing up and saying, "Mr. Lu, we need to go outside to discuss something."
Lu Mingzhou nodded slightly, and the group left.
The founding members of Ximing were still seated when Lu Mingzhou smiled at them, "President Xue, please guide us in the future."
Xue Zhaoheng seemed to have just come to his senses, gave Lu Mingzhou a slight nod, and said nothing.
Lu Mingzhou knew that he was in a bad mood. The company was not doing well and the investors had changed, so it was normal for him to be in a bad mood.
But his mood quickly improved.
Source Intelligence will naturally have subsequent investment plans, but there's no need to tell the investors. Let's wait until after the negotiations are over and then have a private discussion with the founding team.
After a while, the investors returned and sat down again.
Wang Xian said, "President Lu, 520 million."
Lu Mingzhou remained silent.
Wang Xian continued, "Based on the company's overall valuation of 520 million, the combined transfer of 53.6% by several of us would be worth 278.72 million. This is our lowest price."
Since the investors didn't all enter at the same time, at this trading price, some people will definitely lose money while others will make money.
But in any case, it's better than bankruptcy liquidation.
Lu Mingzhou lowered his eyes and silently calculated in his mind.
Source Intelligence was able to acquire a controlling stake in Ximing, its supply chain, inventory, patents, and engineering team with less than 300 million in cash.
"Can."
The investors on the other side visibly breathed a sigh of relief.
But Lu Mingzhou hadn't finished speaking.
"I have two more conditions."
The person on the other end, who had only managed to breathe halfway through, stopped again.
Lu Mingzhou held up two fingers.
"First, during the due diligence process, if any discrepancies are found between historical accounts receivable, unreceived orders, inventory impairments, and current records, the consideration must be deducted."
"Secondly, core team members cannot leave the company within six months of the transaction."
Both of these are common protective clauses in merger and acquisition agreements, and the investors breathed a sigh of relief.
However, Xue Zhaoheng, the CEO of Ximing who had been completely oblivious throughout, had a sudden thought after hearing the second condition and looked at Lu Mingzhou.
However, because the investors were still present, he said nothing.
Wu Tong took out his computer, filled in the equity ratio and amount on the pre-prepared letter of intent, and printed it out on the spot at the printer next to the meeting room. Lu Mingzhou, Wang Xian, and representatives from several other institutions signed it.
Mergers and acquisitions are a very complex process, and it's impossible to finalize a deal in just one negotiation.
Afterwards, a qualified third-party agency needs to be invited to conduct due diligence, and then a second negotiation on the transaction amount will be conducted based on the results of the due diligence.
This is actually a second chance to negotiate the price.
The price quoted just now was the highest price, not the lowest price; this is an unspoken consensus between both parties.
Therefore, the most difficult step has actually been taken.
However, it will take at least two to three weeks to complete this equity transaction.
After the negotiations concluded and everyone left the meeting room, Wu Tong finally breathed a sigh of relief as they walked out.
This was his first time Add One real merger and acquisition negotiation. During his internship at investment banking, he had only done paperwork. Although the amounts on paper were large, it didn't have the same thrill as face-to-face negotiations.
Mr. Lu looked so cool in the conference room just now.
As the two were walking out, someone approached from behind and called out, "President Lu."
Lu Mingzhou stopped and looked back; it was Xue Zhaoheng, the CEO of Ximing.
"President Xue, is there something you need?"
"President Lu, I'd like to see Boss Han," Xue Zhaoheng said. "Otherwise, I won't sign the agreement promising not to resign."
Lu Mingzhou was a little surprised. Xue Zhaoheng was quite perceptive; he could tell from just one sentence during the negotiation that Boss Han was interested in him.
Lu Mingzhou replied, "If the deal goes through, Ximing and Source Intelligence will be one family. You should naturally meet Boss Han, but Boss Han is very busy lately. I need to ask him when he will have time."