반도체 사이클 '이때' 꺾인다 올해 코스피 반등, 여기까지 | 삼성전자 파운드리, 승부수 통했다! 중소형주 폭락, 지금이 바닥이다? | 박세익 체슬리투자자문 대표 |자이앤트썰

Finance-focused summary of the subtitles

Market / macro context & outlook

• The discussion highlights extreme volatility in Korea’s KOSPI/KOSDAQ, including circuit breakers for two consecutive days, followed by a sharp rebound.
• They connect the selloff to Fed policy changes, mentioning:
  • The July FMC meeting, and
  • A “change in the Fed Chair” leading to rate hikes.
• A correction of about -15% to -20% is referenced for July–September, but the dialogue says the actual drop ended up being only ~4% from the peak—less severe than feared.

Semiconductor cycle thesis (core argument)

• Core thesis: “This time is not different.” They argue the semiconductor cycle is not a smooth, continuous uptrend.

• They suggest semiconductors are entering a phase where:

  • Supply dynamics dominate (explicitly stated as the #1 factor).
  • AI/HBM expectations are already priced in, making additional upside harder without new catalysts.
  • Future supply increases are the main risk.
• Repeated caution: a “V-shaped” recovery is unlikely in the short term.

Company & sector-specific claims (tickers included)

SK Hynix & Samsung Electronics

• They claim SK Hynix and Samsung Electronics “doubled” around a product/listing event, which allegedly shaped retail behavior (buying at peaks, selling at lows).
• They argue prior recovery attempts may be difficult, and describe these stocks as being “in the box” for about two years.
  • SK Hynix is described as a “pure semiconductor” company, so they expect a range-bound ~2 years.
• Cycle turn timeframe (their estimate):
  • Stocks are expected to resume rising in 2H 2028,
  • After reflecting new supply in 2027 and beyond.

Micron

• Micron is referenced as an example of how “perception” drives price action, including through the lens of AI/semiconductor expectations.

Changxin Memory (China)

• Changxin Memory (also referenced with variant spelling as Changxin / Changxin Memory) is presented as a major supply shock risk.
• They mention foreign buying of about 46 trillion won related to Chinese board activity after an IPO, described as the “Shanghai Gaochang Board.”

Yangtze Memory / Yangche

• Yangtze Memory / Yangche is described as an aggressive investor/producer that generates profit and expands IP, creating another supply competitor risk.

SanDisk & Kioxia

• SanDisk and Kioxia are mentioned as part of where the next stage of market expansion/adjustment may occur.
• They expect stock prices there to fall materially (see drawdown estimates below).

Explicit numbers & valuation / return metrics cited

• Correction range (July–September): -15% to -20%
• This time’s drop vs peak: ~4%
• Valuation thresholds / re-rating pace
  • Korea PBR < 0.9 is mentioned as a condition that historically hasn’t “exploded within a year” (i.e., no rapid massive re-rate expected).
  • Historical note: when Hynix P/E fell to ~3x, it was viewed as a sell point in prior cycles (“sell when P/E is low for cyclicals”).
  • At prior peaks, Samsung / SK Hynix / Micron did not exceed 10x valuation (as stated in the discussion).
• Profitability swing
  • They suggest margins/operating profit could fall -50% to -70% YoY before a stronger rebound phase.
• Expected drawdowns
  • For Chinese firms and companies like SanDisk and Kioxia, they expect stocks to drop about 70% to 80% from here, framed as a new supply-driven stage.

Time-based stock behavior & market levels

• They claim the TSMC / Samsung / AI infrastructure cycle stage is around Stage 2 or 3 of 9 (as asserted by “those with deeper understanding,” not independently sourced).
• For KOSPI peaks:
  • 9,300 is described as a difficult level to break for the next six months.
• Behavioral timing around events
  • The market may waver two weeks before the FMC due to rate-hike noise.
  • They reference the next FMC meeting in September.
  • Chuseok (early–mid September) is described as a liquidity inflection.
  • Recommendation-like guidance: buy semiconductors before Chuseok, despite rate-hike uncertainty (because expectations can flip quickly).

Framework elements & positioning mindset

• Behavioral/positioning approach (implied):
  • Understand cycle risk and use a contrarian mindset (buy when down, not chasing peaks).
  • Time decisions around macro catalysts, especially FMC and liquidity changes.
• Short-term handling:
  • For roughly 3 months, they recommend managing volatility with “boxes” (range/position discipline) rather than expecting a clean rebound.
• Leverage guidance:
  • “Leverage this time is completely useless.”
• Psychological risk management:
  • Emphasis on enduring drawdowns because markets may not rebound in a clean V-shape.

Investing strategies & portfolio construction cues

• Strategies explicitly named:
  • Buy-and-hold (trend-independent).
  • Contrarian investing: “sell when price goes up, buy when it goes down.”
• Portfolio/vehicle approach:
  • They use ETFs to avoid direct conflicts of interest between “company vs client” (stated operational reason).
• ETF examples mentioned:
  • KODEX 200 (tracks KOSPI 200)
  • A “Tiger” product tracking KOSDAQ 150 (per subtitle)
• Retail positioning suggestion (with explicit numbers):
  • If you have spare funds: buy around 6,000 and reduce position between 7,300 and 8,300 (the subtitle context is not fully clear which instrument strictly corresponds to those levels).

Performance / earnings timing claims

• Expected semiconductor bounce pattern around earnings:
  • Hynix “missed” because high-priced volume was postponed to the next quarter—therefore the “next stumble is good.”
  • They expect Samsung Electronics preliminary earnings in October, with the stock expected to rise sharply ~2–3 weeks before the announcement.
• Supply agreement structure claim:
  • They also discuss the idea that fixed transaction prices may rise as part of the supply/demand agreement (“fixed transaction price/guard”).

Geopolitical risk thesis (foundry & geopolitics)

• Geopolitical risk is framed as central:
  • Taiwan risk by 2027 is mentioned (“annex Taiwan in 2027”).
  • TSMC is described as critical to U.S. AI chip supply; if disrupted, the U.S. would need an alternative foundry.
• Policy-driven stance (strongly implied):
  • The U.S. should foster Samsung foundry.
• Hypothetical numeric scenario:
  • If Samsung foundry share rises from 7% to 17% while TSMC falls from 70% to 60%, Samsung could gain value.
  • They estimate TSMC market cap at 400–500 trillion won per 10% change; thus a 10% share gain could imply ~+500 trillion won.
• Mitigation actions mentioned:
  • TSMC building factories in Japan and the U.S. to reduce tangible risk.
  • Samsung is also said to be acting similarly.

Disclosures / disclaimers

• No explicit “not financial advice” disclaimer appears in the provided subtitles.

Presenters / sources mentioned

• Hong Soon (홍순) — Maekyung Giant (맥경자이앤트썰)
• Park Sik / Park Sik (박세익) — Chaesley Investment Advisory (체슬리투자자문), referenced as Team Leader / CEO
• Also mentioned in dialogue context:
  • Chairman Choi Tae-won (최태원)
  • Senior Hong Jang-won (홍장원)