The Yuseong Group Chairman (1)
Honestly, there wasn’t anyone who understood the semiconductor market outlook for the next ten years better than I did.
Instead of answering at once, I asked, “Don’t a lot of experts within the group and the researchers at the Yuseong Economic Research Institute offer their opinions?”
The Yuseong Economic Research Institute was the largest economics think tank in the country. It produced all kinds of reports on the Korean economy and industry. I had read enough of them to last a lifetime when I worked at a securities firm.
Chairman Yoo Jae-ho smiled at my words.
“An internal perspective and an external one are never the same. The most dangerous moment in business is when everyone speaks with a single voice.”
He was right.
If you were a CEO, you needed to hear a variety of voices.
When someone important tells you to speak freely and not worry about appearances, you had better be careful. If you really did speak without worrying about appearances, you could easily end up out in the cold.
But since I was an outsider anyway, I decided to speak without worrying about appearances at all.
“Yuseong Electronics is the strongest company in the world in memory semiconductors. But memory chips account for only thirty percent of the entire semiconductor market. System semiconductors have the advantage of higher margins and steadier demand. That’s why I think you should be moving aggressively into that market.”
The semiconductor industry could be divided into two broad categories.
One was memory, and the other was non-memory. Everything other than DRAM and NAND fell under non-memory semiconductors.
That included central processing units, or CPUs, used in computers and servers, graphics processing units, or GPUs, as well as camera image sensors, or CIS, used in smartphones and self-driving cars.
Chairman Yoo nodded.
“That investment is already under way in the foundry business.”
To grow the semiconductor industry, you needed technology, of course, but you also needed capital measured in trillions—tens of trillions—of won.
Fortunately, Yuseong Electronics was earning enormous profits from memory semiconductors and smartphones.
Using that foundation, Yuseong Electronics had announced that it would invest 100 trillion won in fostering its system semiconductor business over the coming years.
The investment already under way had presumably been decided after cold, hard demand forecasts from experts. The investment alone came to a staggering 40 trillion won.
How many companies in the world even had the technology and capital to carry out capital spending on that scale?
It was only possible because it was Yuseong Electronics.
I cut straight to the point. “It needs to be doubled from here.”
My words made Sung Yoon-ah look startled, and Chairman Yoo Jae-ho became visibly interested.
With a faint smile, he asked, “Why do you think that?”
“Because everything in the future will become digitized.”
That was something most experts were already predicting.
The problem was speed. The pace of digitalization would move far faster than anyone expected.
People had been talking about a semiconductor supercycle for the past three years, but this was still only the beginning. In a few years, the market would go completely insane.
In effect, every product that ran on electricity would end up using semiconductors, and semiconductors would become scarce.
Car factories would shut down because they couldn’t get chips, and game consoles wouldn’t even be produced.
“It’s unreasonable to keep ramping up capex blindly when you can’t know how many client companies you’ll be able to secure. Besides, the foundry market already has powerful players in it.”
Semiconductor companies could be divided into three broad types.
Companies that only designed chips, companies that only manufactured them, and companies that did both.
Those were called fabless, foundry, and integrated device manufacturers, or IDMs.
Because Yuseong Electronics handled both design and manufacturing, it fell into the IDM category.
The company Chairman Yoo was talking about was none other than PSMC in Taiwan.
It was the world’s largest foundry company, with the best technology in the world. If Yuseong Electronics entered the foundry market, it would inevitably end up going head-to-head with PSMC.
Especially below the 10nm process node, the market was effectively a duel between PSMC and Yuseong Electronics. Even then, PSMC still held the edge at roughly seven to three.
“It’s true that Yuseong Electronics has a fatal weakness when it comes to direct competition with PSMC.”
He must have already known that, but Chairman Yoo asked as though he didn’t.
“And what weakness is that?”
“PSMC’s motto is that it doesn’t compete with its customers. But Yuseong Electronics can’t say the same.”
The biggest fear for any company that outsourced manufacturing was that its own designs or technology might leak.
That was why PSMC had declared that it would not do in-house design, and that had earned the trust of its clients. Giant fabless companies like Enple, Qualcomm, and Endovia were happy to trust PSMC with their production.
But Yuseong Electronics was an IDM.
There might be various safeguards in place to protect client designs, but there was no guarantee that the know-how Yuseong Electronics gained in the production process wouldn’t be applied to its own designs.
From a client’s perspective, that would leave a bad taste in the mouth no matter what. The feeling would be even worse if the client happened to compete directly with Yuseong Electronics.
Chairman Yoo gave a wry smile.
“That doesn’t mean we can give up our own design capabilities, does it?”
I nodded. “I think we need to turn that weakness into a strength.”
“How?”
“By taking on design work and support tailored to what clients need. Yuseong Electronics already has considerable expertise in design.”
Memory semiconductors were standard components used across electronic devices. Non-memory semiconductors, on the other hand, changed in design depending on whether they were going into a PC, laptop, smartphone, or vehicle.
So if memory chips were mass production of a few standardized product lines, system semiconductors were low-volume production across a huge variety of products.
The weakness of handling both production and design could be transformed into the strength of offering design support, or even taking over design work for clients.
Chairman Yoo smiled and nodded. “Hoh. A nice shift in perspective.”
Even so, he didn’t seem all that surprised.
That was only natural, because what I was saying was already in motion. It wasn’t as though Yuseong Electronics lacked smart people.
In fact, not only Yuseong Electronics, but even Antel, which would later jump into the foundry race, would adopt the same approach.
“To do that, you’ll need to strengthen your design capabilities beyond where they are now.”
“So you’re saying we should increase our investment in design?”
Manufacturing tended to deliver results in proportion to the money you spent. Design, however, was not like that. There were plenty of cases where you poured money into it and still failed.
The best way to solve that was...
“You need to actively acquire fabless companies.”
Chairman Yoo shook his head.
“There are too many bad ones, and the good ones are already overvalued by the market.”
When people heard “fabless,” it sounded grand, but in reality there were fabless companies everywhere.
The semiconductor industry was one where a single genius could do the work of ten thousand people.
Because no factory was needed, the capital requirements were low, and since there were companies that provided foundational designs, fabless firms sprang up like weeds.
But among all those companies, how many actually made a profit?
On the other hand, the ones that did make money already commanded absurd valuations.
“And besides, those kinds of companies aren’t something you can just buy because you’ve got the cash.”
Semiconductors were a strategic national industry.
They were tied directly to national security, and because monopoly concerns were always in play, approval was needed not only from the country in question but often from related nations as well.
In fact, deals were often derailed right before completion because of exactly that sort of issue.
“If you look among startups, though, there are plenty of promising companies. A small business with just a handful of employees can grow into something huge later on. You need to identify the good ones and acquire them.”
In that case, there was no need to go through a complicated approval process.
Chairman Yoo said, almost jokingly, “I suppose I just don’t have that kind of talent.”
Sung Yoon-ah was startled and waved her hands in protest. “Ah, n-no, Chairman. That’s not what I meant at all.”
But the truth was, he wasn’t wrong.
If he had been better at it, Yuseong Electronics wouldn’t have ended up getting tripped up later on.
“Even if we want to increase investment in foundries, there are practical limits. Building a semiconductor plant isn’t enough by itself.”
I nodded. “You need the equipment and the manpower to run it.”
“Especially with EUV equipment, production is limited. Even if you build the factory, it’s useless if the EUV tools don’t come in.”
The equipment inside a semiconductor fab cost sums that defied imagination. Among them, the most expensive was the EUV tool used in lithography.
The price was beyond outrageous.
A recent model was said to cost well over 150 million dollars apiece.
It was that expensive not only because it was extraordinarily difficult to make, but because there was only one company in the world that produced it.
That company was ESML in the Netherlands.
The bigger problem than price was that production was limited. The manufacturing process was so difficult that it was hard to supply the tools on time.
Yuseong Electronics was the supreme power broker in semiconductors, so it wasn’t used to begging anyone for anything. But ESML was the one exception.
In fact, a few years later Yuseong Electronics would finish building its factories and still be unable to bring in the EUV tools, leaving the plants unable to operate.
Eventually, companies would even start outbidding one another just to get their hands on the equipment.
It was time to bring up the real point.
“You need to find an alternative to EUV.”
“How do you mean?”
“The most reliable way is to acquire a company that has the related technology.”
“Is there such a company?”
Only then did Sung Yoon-ah seem to realize what I was getting at, and she looked at me with a startled expression.
I went on without caring. “Yes. There is.”
Chairman Yoo showed interest. “What company?”
“Dongwoo Precision.”
“The one that was recently suspended from trading.”
Dongwoo Precision was one of the larger semiconductor equipment companies. It made sense that he’d know it.
“Do you happen to have some connection to that company?”
I answered honestly. “Actually, I bought its corporate bonds with a friend.”
Chairman Yoo burst into amused laughter.
“Ha ha ha! So what you’re saying is that you want me to pay off your debts for you?”
He said it like a joke, but there was an edge hidden in the words. If I said the wrong thing here, would I get kicked out?
But I didn’t hedge.
“That’s right.”
“Miru!”
Sung Yoon-ah cried out in shock, and Chairman Yoo asked, “And why should I do that?”
I gave him the answer in a single sentence.
“Fifty billion dollars.”
“Pardon?”
“The technology Dongwoo Precision holds is worth more than fifty billion dollars. So Yuseong Electronics needs to acquire it before any other company steps in.”
Right now, it was something only I knew, but XRT Semicon’s acquisition of Dongwoo Precision would later be called a disaster for Korea’s semiconductor industry.
The reason was that behind XRT Semicon stood CMIC, China’s largest semiconductor foundry company.
Reminder: All system notifications, comments, letters, status windows, and other special in-universe content must use styled div blocks with inline CSS from the templates. Never output them as plain p text.