Ideas
Hyperscaler worries overdone; M7 undervalued.
Hyper-scalers are not the main victims of the AI pace-adjustment debate; their capex commitments remain intact, Oracle's operating cash flow improvement and falling CDS premium show financial conditions are better than feared, and M7/hyperscaler valuations look undervalued.
AI pace talk pressures memory near term.
The AI labs' pace-adjustment talk is likely a short-term negative for memory semiconductors and AI infrastructure because frontier-model token-spend growth could slow and compress multiples, even though hyperscaler capex commitments mean earnings and prices are unlikely to collapse over the next 6-12 months.
Buy KOSPI dips for October.
September macro noise and AI pace-adjustment fears are creating a temporary KOSPI decline, but the downside below the mid-6,500 area is limited; investors should not sell into weakness and should buy dips for a better October market.
Oil likely falls below $100.
Oil above $100 is historically rare and rarely sustained for more than a month; with the Middle East disruption already reflected, oil is more likely to fall back below $100 in October, easing inflation and rate-hike pressure.
Hardware innovation cycle accelerates on cost.
Rising memory and display prices are not ending the IT hardware cycle; they are forcing hardware makers to innovate faster, pulling forward foldables, smart glasses, humanoid robots and other premium devices, so the hardware upgrade cycle should accelerate into 2027.
HBM leadership diversifies beyond SK hynix.
HBM supply is diversifying into 2027 as Samsung Electronics and Micron expand capacity; this should narrow or reverse SK hynix's HBM share and removes the prior rationale for SK hynix to outperform Samsung/Micron so dramatically.
HBM leadership diversifies beyond SK hynix.
As non-cyclical HBM revenue mix rises, Samsung Electronics' forward PBR should find a bottom near 1.5x; a move toward the 2x mid-band could imply roughly 30% 12-month upside, making the stock a steadier memory holding than SK hynix.
HBM leadership diversifies beyond SK hynix.
HBM supply is diversifying into 2027 as Samsung Electronics and Micron expand capacity; this should narrow or reverse SK hynix's HBM share and removes the prior rationale for SK hynix to outperform Samsung/Micron so dramatically.
KOSDAQ oversold; liquidity rotation likely.
KOSDAQ has been overly punished, holding around 800 despite high rates; as SK hynix's share of Korean trading volume falls, liquidity should rotate toward KOSDAQ and stock-specific small/mid-cap themes.
Diversify into AI capex Q-cycle.
2027 is when AI data-center plans move into execution, creating a Q-cycle that favors non-memory beneficiaries; portfolios concentrated in memory should diversify into construction, semiconductor equipment/materials, and raw materials as capex and commodity demand broaden.
Solar near term, nuclear long term.
Data-center energy demand must expand several-fold by 2040; solar is the only readily available energy source near term, while nuclear is the fundamental solution but takes about 15 years, so 2027 should bring policy and project momentum for nuclear/SMR.
Solar near term, nuclear long term.
Data-center energy demand must expand several-fold by 2040; solar is the only readily available energy source near term, while nuclear is the fundamental solution but takes about 15 years, so 2027 should bring policy and project momentum for nuclear/SMR.
ESS orders and earnings visible by 2027.
ESS is approaching visible order and earnings recognition in 2027; CATL's ESS sales remain strong, while in Korea Samsung SDI is a purer ESS play and LG Energy Solution offers greater capacity/order scale but less purity, so both are decoupling from EV battery materials.
LFP conversion could lift battery materials.
Korean battery material companies are mostly exposed to ternary EV batteries and have lagged; they could outperform when they mass-produce or convert to LFP, which is increasingly favored by ESS demand.
HD HHI engine/SMR investment rerating.
HD Hyundai Heavy Industries' 1 trillion won investment in medium-speed engines/SMR should expand its energy equipment capacity from 3GW to over 7GW, surpassing Wärtsilä; the stock trades at a discount to Wärtsilä and could see a valuation catch-up and market-cap increase larger than the investment amount.
Shipbuilders need Hormuz; prefer engines.
The broader shipbuilding cycle still needs the Hormuz Strait to reopen, and rising engine prices will squeeze margins for yards without engine capability; investors should focus on ship-engine exposure rather than the whole shipbuilding sector.
This 3PRO TV (삼프로TV) video, published September 14, 2026,
features Lee Hyuk-jin
discussing M7, ORCL, SMH, EWY, WTI, Humanoid robots, XTH, MU, 005930.KS, 000660.KS, KOSDAQ, Korean construction, Korean semiconductor equipment/materials, DBC, SOLAR, NLR, 373220.KS, 006400.KS, CATL, LFP battery materials, 329180.KS, Korean shipbuilding sector.
16 trade ideas extracted by AI with direction and confidence scoring.
Speakers:
Lee Hyuk-jin
· Tickers:
M7,
ORCL,
SMH,
EWY,
WTI,
Humanoid robots,
XTH,
MU,
005930.KS,
000660.KS,
KOSDAQ,
Korean construction,
Korean semiconductor equipment/materials,
DBC,
SOLAR,
NLR,
373220.KS,
006400.KS,
CATL,
LFP battery materials,
329180.KS,
Korean shipbuilding sector
