중국 메모리의 무서운 성장, 3년만에 1%가 10%됐다 (박종훈의 지식한방)

Finance/Markets-focused Summary (Semiconductors, China Competitive Risk)

Main Claims / Market Context

• The speaker argues that Korean semiconductor momentum is strong, but warns against underestimating China—particularly DRAM producer CXMT (Changxin Memory Technologies)—which could “burglarize” Korea’s market position.
• Competitive shift described:
  • Korean firms: moving from general-purpose DRAM toward HBM (High Bandwidth Memory).
  • Chinese firms (allegedly): filling the gap in general-purpose DRAM by rapidly scaling CXMT.

Key Tickers / Companies / Entities Mentioned

• CXMT (Changxin Memory Technologies)
• Samsung Electronics
• SK hynix (spelled inconsistently as “SK INTIX” / “SK Inix”)
• Micron
• Huawei
• Goldstar
• NEC
• V-Tech
• Coive (referenced as a data-center builder)
• “Yangju Memory / Yangzhou Memory” (second Chinese memory competitor referenced)
• Goldman Sachs (forecast source)
• Nomura (forecast source)
• Commerce Secretary Rutnick (US official referenced)
• “US Semiconductor Agreement” (historical context)

---

DRAM Market Share / Timelines and Forecasts (Numbers)

CXMT DRAM Market Share

• 2023: 1%
• Q2 2025: 4%
• Q2 2026: 10%
• Claim: CXMT reached the previously forecast 10% by 2028 two years early.
• Forecasts referenced:
  • Goldman Sachs: CXMT could supply/handle 50% of China DRAM demand by 2028.
  • Nomura: CXMT could reach 18% global market share by 2028.
    • The speaker also frames a survival threshold at 16–17%.

Capacity / Wafers

• Speaker claims CXMT plans 350,000 wafers by end of this year.
• Micron usage cited: 385,000 wafers per month.
• Additional scale note: CXMT is “about half the size of Samsung Electronics or SK hynix.”

---

Profitability / Margins (Explicit Percentages)

Margin Comparison (as stated in the narrative)

• Samsung Electronics: 70%
• SK hynix: 76%
• Micron: 80%
• CXMT: 82% (noted as currently higher in the speaker’s framing)

Why CXMT Margins Are Rising (as described)

• Explanation offered:
  • The US/Korea HBM transition allegedly caused DDR5 supply shortfalls.
  • Claimed “regime change”: as of January (this year), HBM profitability became lower than DDR5 profitability (described as counterintuitive).
  • Result: DRAM profits were pulled forward, enabling faster reinvestment and catch-up.

Caution / Disclosure Point

• The speaker claims CXMT’s own disclosure allegedly states the margin rate is “unsustainable.”
• The speaker cautions not to take Chinese claims at face value, citing US oversight/monitoring.

---

Investment / Competition “Mechanism” Described

“Chicken Game” Logic

• Companies ramp HBM production aggressively because whoever dominates advanced processes shapes future leadership.
• This allegedly creates a DRAM production gap, helping latecomers like CXMT sustain higher DRAM margins.

China’s Industrial Policy Support (as claimed)

• CXMT allegedly benefited from:
  • subsidies covering sales drops,
  • higher purchase prices from domestic buyers (state/major clients),
  • state-backed financing and procurement,
  • rapid scaling through equipment and funding.

“Unlimited Capital” Risk

• If China escalates and continues funding losses, competitive pressure on Korea could intensify.

---

HBM Supply/Demand and “Worst Year” Timeline

• The speaker states 2027 is expected to be the worst year for supply (HBM-related supply bottlenecks), which could benefit Korea.
• However, if that also benefits China’s DRAM/HBM trajectory, CXMT could sustain or increase profit margins while catching up.

Wafer Commitments (as stated)

• Micron HBM wafers:
  • 40,000–50,000 wafers/month
  • targeting 100,000 by end of this year
• Speaker warning: this may increase pressure across broader DRAM supply.

---

Interest Rates / Financing Risk (Macro Finance Element)

Financing-Dependent Demand Risk

• The speaker argues latecomer/demand outcomes depend on financing conditions.
• Example provided: HBM data-center builders (e.g., Coive) are borrowing.
• Claim: if interest rates rise, demand could fall even if the AI cycle continues.

Rate Thresholds Mentioned

• Hold out until interest rates reach around 4.5%
• If 5%–6%, the interest burden becomes “huge”
• Corporate bond maturities in 2027 mentioned:
  • implies rollover costs may jump for 3–5 year bonds issued earlier at low rates.

---

Scenario Framework (Implied Outcomes)

1. HBM demand stays strong AND CXMT fails to grow
   • Korea’s excess profits persist / the boom continues.
2. HBM growth slows (e.g., due to higher rates)
   • the sector becomes cyclical again → risk of a “memory recession.”
3. CXMT captures ~50% of China DRAM market by ~2028 AND HBM demand stays strong
   • Korea may still profit from HBM,
   • but could lose the general-purpose DRAM replacement market, limiting upside after HBM profitability fades.

Survival / Market Share Threshold

• ~16–17% global share described as a threshold for surviving in memory.

---

Recommendations / Cautions (Explicit)

• The speaker emphasizes that Korea must not become complacent.
• Korea should treat China as a national-level competitive threat.
• “Secure ammunition” is framed as building cash/financing capacity to outcompete.

Government / Infrastructure Priorities Emphasized

• Build power supply capacity (risk of power shortages).
• Construct transmission towers (risk that power connectivity for fabs may fail if buildout lags ~20 years).
• Invest in talent development and retention (risk of AI talent moving to the US/Europe).
• Protect trade secrets / cybersecurity (risk of hacks; claim: China is “2–3 generations behind,” but one major breach could be fatal).

Internal Policy Warning

• Criticism that some in Korea may try to “kill the goose” (reduce reinvestment/oversight) because of current profits.
• The speaker argues this could damage future competitiveness.

---

Tickers / Assets / Instruments Mentioned (Capital Markets)

• No specific tradable tickers or ETFs are named.
• The discussion focuses on interest rates and corporate bonds (financing risk).
• No explicit yields/spread numbers provided beyond 4.5% / 5% / 6% thresholds.

---

Disclosures / Disclaimers

“I am absolutely not telling you to talk about stock prices today.”

• A stock-price disclaimer is included.
• No explicit “not financial advice” disclaimer is stated, beyond that stock-price caution.

---

Presenters / Sources Mentioned

• Park Jong-eun (referenced via the video title: “박종훈의 지식한방”; subtitles show inconsistent naming as “박종훈” / “박종훈”)
• Goldman Sachs
• Nomura
• US Commerce Secretary Rutnick (“Rutnick”)
• Historical reference: “US-Semiconductor Agreement” (no individual named)