Chapter 177: Premium

Equipment-related discussions had dragged on, and time had slipped by unnoticed.

However, the original agenda items remained largely untouched.

Oh Seong-hak quickly moved on to the next item.

“Next are the investment items outside of semiconductor equipment.”

[Dojin Group Investment Details]

  1. Semiconductor equipment development

…

…

…

  1. Quantum 4-S, A smartphone development

  • Development lead: Hwang Chilbong, Dojin Tech

  • Development cost: 3.2 trillion won

  1. Neuromorphic project

  • Development lead: Hwang Chilbong, Dojin Tech

  • Development cost: 32 trillion + α

  1. Luna platform development

  • Development lead: Han Yu-ju, team leader

  • Development cost: 5.5 trillion won

  1. Others

  • Global data center operations and infrastructure expansion

  • Development of smart TVs, washing machines, refrigerators, etc.

  • Advanced technology convergence research institute funding

  • AI research talent acquisition and industry-academia collaboration project funding

  • High-bandwidth memory semiconductors (HBM)

  • Dojin Bio……

-……

The room fell silent as the figures appeared on the screen.

“……Adding up those figures, it’s definitely over 100 trillion, right?”

“Wow, no matter how much money we make, it really never ends.”

“But is there a reason to push for this level of investment, even if it means overstretching ourselves?”

I replied carefully to Hong Chung-ki’s question.

“It’s not overstretching, but rather pre-emptive preparation to seize the next opportunity.”

“By opportunity, you mean… the AI-related business you’re pursuing, sir?”

“Exactly. Just as feature phones were replaced by smartphones, I believe the future of the internet will be reshaped by whoever dominates AI.”

“Is it… that significant?”

Hong Chung-ki was visibly surprised, but others in the room were equally stunned.

“Starting with lenses, then smartphones, followed by our proprietary OS Lunar and KTalk, and now KTalk Feed—all of this is merely preparation for the AI-driven world.”

The atmosphere in the room grew heavier.

It sounded less like a simple business strategy and more like ‘intentional evolution.’

But my words were true.

Now that we had a solid foundation, every step was critical.

One misstep could jeopardize my plan to dismantle Ahsung’s core semiconductor business and dominate the AI market.

“So, for the time being, we’ll be funneling all our profits into AI, as you’ve outlined, sir?”

“Yes… that’s right.”

Oh Seong-hak’s comment lightened the mood somewhat.

“But I understand everything else now. What exactly is the Neuromorphic project that Hwang Chilbong is working on?”

Chilbong stepped in to answer Park Cheol-jin’s question.

“The Neuromorphic project aims to develop the next-generation computational architecture for the Luna AI platform.”

Chilbong, now holding the mouse, displayed circuit diagrams and the keyword ‘Neuromorphic Computing’ on the screen.

“Initially, the Luna platform will use parallel computing chips like GPUs and TPUs.”

In reality, Chilbong was developing a microarchitecture called Black Robin under my guidance.

“Then… why are we developing Neuromorphic?”

“While the chips for Luna remain the same, Neuromorphic performs computations by mimicking the human brain—specifically, the structure of neurons and synapses.”

Together, Chilbong and I aimed to implement a non-linear, non-reversible computational model, moving beyond traditional transistor-based systems.

“……”

Park Cheol-jin and others looked perplexed, as expected.

I smiled and added an explanation.

“Simply put, think of it as a chip that performs computations in a way much closer to human thinking than the Black Robin chip currently in development.”

Hong Chung-ki nodded at my explanation.

“If GPUs process commands and data in a fast, serial manner, Neuromorphic spreads and accepts information simultaneously, like a neural network.”

Despite his background in business administration, his quick understanding was impressive, likely due to his master’s degree from Korea University.

“As Hong said, the human brain doesn’t collapse due to a single faulty neuron. Neuromorphic aims for both structural flexibility and energy efficiency.”

Chilbong added,

“The GPUs currently in production or development at Ensedia are highly performant but consume a lot of energy and require excessive resources to maintain parallel structures. The same goes for our Black Robin.”

Neuromorphic promised high inference efficiency with minimal energy consumption.

“So, that’s why the development cost is 32 trillion?”

“Even 32 trillion is a conservative estimate. Once development ramps up, costs could multiply.”

Even LUNA-EVX was just foundational technology for Neuromorphic production.

“Wow, so how much more money are we talking about?”

Seong-hwan, who had been quietly listening, couldn’t hold back.

“There won’t be a moment to spare cash for at least a few years.”

Everyone’s mouths dropped open.

Oh Seong-hak sighed deeply once more.

“……Still, most projects at Dojin Tech, Dojin Soft, and Dojin Frontier are progressing steadily. If we maintain this momentum, I don’t foresee major issues.”

“But… sir, if anything goes wrong, even a minor setback, we could face liquidity problems.”

Such an event seemed unlikely, but if it did occur, the consequences could be dire.

＊＊＊

Mujinwon.

Chairman Cheon Ho-tae gazed silently out the window.

The seasons had shifted from spring to summer, and the warm breeze was palpable even through the gray glass.

Soon, a cautious knock sounded, and Kim Seong-jae, Head of Strategic Planning, entered.

“Chairman, I’ve come to report as you requested.”

Without turning, Cheon Ho-tae asked,

“Is Jun-young still in Hong Kong?”

“Yes, since his transfer to the Hong Kong branch, he’s been keeping a low profile without causing any trouble.”

“Tsk. At least he knows his mistake.”

The Titan battery explosion had forced us to withdraw from the smartphone business.

It was a necessary move to show the board that action was being taken.

This complicated Cheon Ho-tae’s succession plans.

“What about Jun-seok?”

“The handover within the semiconductor division’s operational lines is complete.”

“Then officially appoint him as division head this week without delay.”

Kim Seong-jae nodded.

He paused, as if gathering his thoughts, before speaking slowly.

“I have something related to report.”

At Kim Seong-jae’s serious tone, Cheon Ho-tae finally turned and walked to the conference sofa.

As he sat down, Kim Seong-jae placed his A-Pad on the table and began his briefing.

“Recently, we’ve detected semiconductor-related movements at Dojin Tech. While they’re keeping it confidential internally, our analysis suggests they’ve begun designing their own AP and memory semiconductors.”

“AP and memory semiconductors?”

“It seems they plan to use them in the next Quantum model.”

Cheon Ho-tae smirked.

“Do they think it’s competitive?”

“I was curious too, so I dug deeper. Besides the AP, they appear to be developing high-bandwidth memory semiconductors, or HBM, not standard memory.”

“HBM? That’s used in data centers or GPUs. We’ve confirmed it’s not very profitable. They’re developing that? For smartphones?”

Cheon Ho-tae’s eyebrows twitched in disbelief.

“Internally, they believe Kang Tae-joon is grasping at straws.”

Ahsung Electronics had previously produced HBM for Ensedia and Nentendo, but high production costs and unstable yields led to a full business reassessment.

“Aren’t the LPDDR4 currently in production and the LPDDR5 in development sufficient for smartphone performance needs?”

“More than sufficient, actually.”

“Then why use HBM, which consumes more power and poses overheating risks?”

“Perhaps because we dominate the DDR memory market, they’re seeking opportunities elsewhere.”

Cheon Ho-tae scoffed again.

“You never know how much it hurts until you experience it.”

“From what we’ve gathered, they’re pouring most of their Quantum sales profits into HBM and AP development.”

“Is that so?”

Cheon Ho-tae seemed intrigued.

“Including factory construction, equipment purchases, development, and semiconductor design costs, they’ve likely spent over 30 trillion won, with more to come.”

“So, despite earning massive profits from Quantum sales, their liquidity is tight.”

Cheon Ho-tae twisted his lips into a smirk.

“This could be our chance. With the recent WSPI project and their constant interference, it might be best to remove them now.”

Cheon Ho-tae nodded in agreement.

“What approach would you suggest?”

“I think we should leverage the equipment they’re trying to purchase.”

“Equipment?”

Noticing Cheon Ho-tae’s curiosity, Kim Seong-jae elaborated.

“Dojin is looking to buy key equipment like exposure machines, deposition machines, TSV (Through Silicon Via) equipment, and high-speed aligners for HBM.”

Kim Seong-jae paused to catch his breath before continuing.

“Among these, exposure machines, CMPs (chemical mechanical polishers), and high-speed aligners are backordered by over a year.”

“……So?”

“By using our influence to pre-order this equipment, we can delay Dojin’s deliveries. If they want it faster, they’ll have to pay a premium.”

At the mention of a premium, Cheon Ho-tae slapped his knee.

“Let’s highlight their liquidity issues.”

“Exactly. And additionally…”

Kim Seong-jae tapped his tablet a few times to display extra data.

The screen showed a list of critical components Dojin was co-developing with major overseas firms.

“If we halt the extreme ultraviolet mirror development with Germany’s Zeiss, listed here, the impact will be greater than the equipment delay. Also…”

Kim Seong-jae mentioned disrupting the supply of other core materials.

Cheon Ho-tae nodded, looking satisfied.