Just very helpful timelines reiterated around glass substrate (source: Trendforce):
- SKC Absolics (011790) H2 2026 (first mover x $AMAT) - $AMD customers
- Samsung electromechanics h2 2027 (009150) x Sumitomo Chem (4005) - Apple / $AVGO / hyperscalers
Idk about $INTC 2030 reports, we’ll see.
$TSM CoPoS was 2-3Y was correct though from recent TSM chairman comments. Innolux was interesting beneficiary. $SHMD should be too off TSM but financials were pretty toxic.
Same players should appear multiple times, eg innolux + SKC.
Also applies to $LPK and upstream equipment seller around these ramps.
Just spotted a sharp new signal in AI buildout: iHBM (integrated HBM cooling).
SK Hynix just launched memory with built-in cooling elements (ICE) inside the package, delivering ~30% lower thermal resistance. Positions them strongly for HBM5 and next-gen GPU stability.
It just jumped to #1 in today’s asymmetric signal ranking (highest edge + technical quality). Extremely interesting technically, but frustratingly hard to get clean direct investment exposure right now.
Any ideas?
The realistic float for $HYPE is ~200m tokens
At a pace of $15m/trading day inflows and current $58 price, that will retire around 5-5.5m HYPE/month. The Assistance Fund gets a further 1-1.5m a month
So two tiny ETF issuers and the protocol itself would retire ~3-3.5%/month
👇
If you are looking to get exposure to the AI trade without choosing specific stocks, then here are the ETFs you should consider checking out:
This covers 80-90% of the AI trade.
1. Materials
- $REMX: Tracks companies involved in mining, refining, and processing strategic and rare earth metals (e.g., Neodymium, Dysprosium, Lithium). In the AI trade, this represents the absolute furthest upstream physical bottleneck. These materials are critical for manufacturing the ultra-powerful permanent magnets used in data center cooling pumps, advanced robotics, and the precision hardware architecture of high-performance servers.
- $IYM: Provides exposure to chemical producers, industrial gas suppliers, and basic commodity miners. For AI infrastructure, this layer is vital for supplying high-purity industrial gases (like nitrogen and argon from Air Products/Linde) required in sterile semiconductor fabrication environments, alongside copper (Freeport-McMoRan) for power grid transmission and advanced thermal management materials.
2. Equipment
- $EUV: Concentrates on the gatekeepers of semiconductor manufacturing equipment. This layer captures the highly specialized, wide-moat companies that build the lithography, etching, and metrology machines (like ASML's EUV lithography) without which cutting-edge sub-3nm AI chips cannot physically be printed or packaged.
3. Memory
- $DRAM: Focuses on the global oligopoly dominating volatile and non-volatile memory storage architectures. In AI workloads, standard compute is useless without massive, ultra-fast data transfer. This layer targets the pure-plays delivering High Bandwidth Memory (HBM) and NAND, which act as the critical data pipeline feeding massive datasets directly into AI accelerators.
4. Semiconductor
- $SMH: Captures the core designers, developers, and dominant foundries of front-line processing units. This represents the primary "compute" layer of the trade, featuring the massive design moats (NVIDIA, AMD) creating specialized GPU architectures alongside the primary outsourced foundry (TSMC) that physically manufactures the vast majority of the world's advanced AI silicon.
5. Hyperscalers
- $MAGS: Focuses purely on the mega-cap technology titans that dictate the CapEx cycle of the entire industry. These companies function as the primary buyers of hardware, developers of proprietary foundational LLMs, and landlords of the cloud platforms (Azure, AWS, Google Cloud) that rent out AI compute capacity to enterprise customers.
6. Power
- $NLR: Targets the nuclear energy value chain, spanning fuel miners, reactor components, and independent nuclear power generation utilities. As AI data centers demand unprecedented, non-intermittent (24/7) carbon-free baseload power, this layer tracks the primary beneficiaries signing massive, premium-priced Power Purchase Agreements (PPAs) directly with hyperscalers.
- $GRID: Tracks engineering, construction, and electrical equipment providers responsible for upgrading macro power grids and localized data center substations. This layer covers the industrial bottlenecks required to step down high-voltage power, handle distribution, and supply the massive transformers, switchgears, and thermal liquid cooling systems required to sustain next-generation high-density AI server clusters.
Upcoming ETFs
$LYTE: photonics/optics.
$NCLD: neoclouds
The 10-Yr getting above 4.5% is a sign the economy is overheating. Looking for a cooling period in markets. Maybe a consolidation for a month.
Still bullish assets tied to AI and the build out.
$EWY $SMH
[한국 주식 주도주 정리 - $IBKR 외국인 친구들을 위한 가이드 🇰🇷]
1. 반도체
$MU 가 PER 20배 넘어가는데 SK하이닉스($000660.KS) 아직 12배? 13배정도 되었나 이제?
같은 HBM 칩 만드는데 우편번호만 다름.
Korean = NO.1 반도체. 그냥 사면 됨.
- SK하이닉스 $000660.KS - HBM 세계 1위, 엔비디아 최대 공급사. 설명 끝.
- 삼성전자 $005930.KS - 폰+칩+가전을 $MU 밸류에이션으로 살 수 있음. 이게 말이 됨?
- SK스퀘어 $402340.KS - SK하이닉스 지분 들고 있는 지주사. 하이닉스 고베타픽. 할인된 가격에 하이닉스 사는 느낌.
소형잡주도 좀 적어보자면
- 파두 $440110.KQ - SSD 컨트롤러 만드는데 스페이스X 공급망에 엮여있음. 소형주 로켓.
- 인텍플러스 $064290.KQ - 반도체 외관검사 장비. 인텔 공급망 노출. 곡괭이 삽 플레이.
2. 금융주
$IBKR 이 한국 주식을 전 세계에 열어줬음.
그 통행료 받는 게 누구냐? 한국 증권사들임. $IBKR 이 얘네 대신 돈 벌어주는 구조.
- 삼성증권 $016360.KS - $IBKR 유일한 공식 중개 브로커. 독점임. 끝.
- 유안타증권 $003470.KS - 그 다음 IBKR 수혜주. 대만계 모회사라 글로벌 브릿지.
- 미래에셋증권 $006800.KS - 스페이스X 지분 직접 투자까지 함. 호재 덩어리.
- 키움증권 $039490.KS - 한국 개인투자자 거래량 1위. 한국판 로빈후드인데 수익은 진짜 남.(토스증권이 찐이긴하지만...비상장)
- 추가로 4대 금융지주(KB금융 $105560.KS, 신한지주 $055550.KS, 하나금융지주 $086790.KS, 우리금융지주 $316140.KS, 메리츠금융지주 $138040.KS)도 밸류업 정책 수혜로 결국 좋은 소식 올 것 같음.
3. 광통신
$CPO $LITE $CIEN 좋아하는 너네들 들어봐.
걔네 공급망에 한국 애들이 엮여있는데 아직 아무도 모름.
- 대한광통신 $010170.KQ - 이름이 직관적. Korea Fiber Optics. 더 설명 필요함?
- 오이솔루션 $138080.KQ - OI = Optical Interconnect. 한국어로 읽으면 오이(🥒)임.
- RF머트리얼즈 $327260.KQ - RF/광학 소재 만드는 회사. 재료 파는 쪽이라 마진 좋음.
4. 유리기판
AI 칩이 고도화될수록 실리콘 기판이 한계에 부딪힘.
유리기판이 그 다음 세대인데 한국이 이걸 먼저 잡았음.
코스닥이 이미 난리났고 우리는 아직 초입임.
- SKC $011790.KS - 한국 유리기판 대장주. 이 섹터 얘기할 때 빠지면 안 됨.
- 필옵틱스 $161580.KQ - 유리기판 가공 장비. 삼성 DS 부문이랑 합동 제안서 낸 사이.
- 켐트로닉스 $357780.KQ - 유리 인터포저 소재. 필옵틱스랑 세트로 묶어서 보면 됨.
- 기가비스 $420770.KQ - 유리기판 검사 장비. 칩 하나 만들 때마다 얘 매출 나옴.
5. 반도체 소부장 (소재·부품·장비)
- 한국 소부장 = 아무도 얘기 안 하는 Tier-1 공급사들.
K-반도체 골드러시의 곡괭이랑 삽 파는 애들.
- 테스 $095610.KQ - PECVD/CVD 장비. 삼성전자·SK하이닉스랑 차세대 장비 공동개발 중.
- ISC $095340.KQ - 반도체 테스트 소켓. 칩 만들면 무조건 테스트해야 함. 반복 매출 구조.
6. 지주사 플레이 (밸류업)
한국 정부가 지금 상장사들한테 자사주 소각하고 주주환원 하라고 드라이브 엄청 걸고 있음.
지주사들 NAV 대비 40~60% 할인된 채로 거래되는 애들 많음. 정책이 먹히면 그냥 공짜 수익임.
반도체 + 지주사 1타2피 픽:
- 두산 $000150.KS - 지주사인데 반도체/에너지 사업 리스트럭처링 중. 가치 unlock 기대.
- LS $006260.KS - 지주사인데 LS일렉트릭(전력 인프라) + 소재 사업까지 엮임. 요즘 핫.
전통 지주사 밸류업 후보:
- SK Inc. $034730.KS
- 한화 $000880.KS
- CJ $001040.KS
이정도? 요새 코스피가 난리인데, 한국 주식시장이 외국형들한테 열리는것 같아서 좋은 기회라 생각해서 정리해봤어.
한국 시장에 대해서 관심 있거나 궁금한 주식, 정보, 기업 등 있으면 팔로우 하고 소통하자구.
웰컴 😍
Here's a basket of 10 stocks in different thematics that I like right now:
1. $HL - People aren't bullish enough on silver imo. Everyone is focused on copper (also bullish) but silver probably has more upside. I'm in 3 silver plays right now.
2. $PRCT - Relative to other MedTechs it's cheap though understandable as it hasn't proved profitability yet. Estimates suggest FY27 as EBITDA profitability year in which case I think we get a big re-rate at an inflection point.
3. $AMZN - Run the numbers on it yourself or just search on my page "$AMZN". You'll see my assumptions.
4. $FEIM - This is an indirect play on the space trade and genuinely a very interesting company. They make the atomic clocks and precision timing systems that go inside satellites/military equipment etc.
5. $OUST - This has quickly become a bigger position for me. It's one of the few companies operating across the whole physical AI perception stack and trades at just 6.5x sales for an estimated CAGR above 40% for the next 3+ years (not many companies fit this criteria).
6. $TAN - Only a small position for me so far but I don't see a world where solar isn't part of the answer for domestic energy.
7. $DLO - Offers nice exposure to the EM's whilst also offering some very solid growth rates. Currently trades at $13.88 but my model on 20x EPS in 2030 gives us a $38 stock price.
8. $LEU - The only US uranium enricher into HALEU. Huge moat.
9. $HROW - If management even get to 80% of what they're expecting in Q4 27, then $HROW is vastly undervalued today. Execution risk yes, but a higher risk bet on a potentially very undervalued ophthalmic company.
10. $SEI - Crushed earnings earlier this week. I'm already +100% on my position and yet still trades fairly reasonably. Risk to reward makes a little less sense now but still what I deem to be a very good company.
Basket of 20 European stocks I like right now (and probably still will be in 2030):
1. $NBIS - everyone knows? Premier European neocloud w/ exposure to non-US sovereign AI demand. Imo, a mini US hyperscaler.
2. $RR (Rolls-Royce) - more than just sexy cars…nuclear propulsion + space tech support SMRs for AI power & orbital systems. Huge UK/European govt contracts. New management team is phenomenal.
3. $SIVE - @X’s favourite company rn? InP lasers into the SiPho qualification cycle. Their lasers are pretty crucial to companies like $MRVL, & therefore to hyperscalers downstream.
4. $BESI - hybrid bonding qualified at $TSM, Samsung & $INTC for HBM4 and 3D logic. Hyperscaler orders + optical transceiver tailwinds = outsized European semi-equipment leverage.
5. $ASML - GOATed European company of the last decade? EUV lithography monopoly - required for every advanced AI logic chip manufactured worldwide…even though $TSM aren’t buying their new machines rn.
6. $ARM - their CPU IP licensing model powers the most efficient AI inference at the edge & in data centers.
7. $NOK - I’m loving the Nokia renaissance atm. Re-rating as a cloud infra leader w/ ~3x order growth from hyperscalers for AI networking. EU industrial & defense contracts compound through the decade.
8. $AIXA - their MOCVD equipment dominates compound-semis production for 800G & 1.6T optical transceivers + photonics.
9. $SPOT - share price is kinda in the gutter rn, but wow, what a company. No-one’s able to overcome Spotify, especially as they’ll probably win on the AI-Audio front.
10. $IQE - one of my favourites. Their epitaxial wafers are the key material for photonics & VCSEL lasers. InP/GaAs leadership makes them a critical enabler in the AI supercycle. $MTSI investment solidifies them for the future.
11. $SMHN (SUSS MicroTec) - their imprint lithography & bonders enable next-gen AI packaging + photonics. Sole-source temporary bonders for CoWoS-L; capacity-constrained through 2027 against $TSM advanced packaging build-out.
12. $RPI (Raspberry Pi) - cos it tastes great. But no, they do cheap single-board computers & semis power edge-AI + robotics prototyping. AI-agent demand is/will drive them. Recent secondary offering anchored by $ARM turns them into a “legit” company for the future.
13. $IFX (Infineon) - their power management semis & analog chips are in every AI server, EV and industrial robot. Their AI power segment forecasted to grow >25% annually through 2030.
14. $M7U (Nynomic) - LayTec subsidiary sits at a crucial chokepoint in InP/GaN MOCVD via $AIXA ramp. Basically a European chokepoint in yield optimization.
15. $ABBN (ABB) - ABB’s electrification, robotics, and motion portfolio directly benefits from AI-driven factory automation and humanoid deployment. Factory automation is already here…and maybe we’ll all have a humanoid bestie at some point in the future?
16. $SOI (Soitec) - their SOI & SmartSiC substrates are the enabling material for power-efficient AI chips and photonics. AI-driven demand for 300mm SOI capacity has already made it Europe’s top-performing large-cap stock of 2026 w/ multi-year ramps locked in.
17. $SU (Schneider Electric) - cleanest direct beneficiary of $2T+ global data center capex through 2030; switchgear/UPS backlog at record. Power management and liquid-cooling infrastructure are the rack-scale backbone for European AI data centers. Grid-modernization + hyperscaler co-location deals compound multi-decade visibility.
18. $LPK - yes, despite my “bear thesis” lol. Glass substrates are a huge bottleneck due to shortages. Their laser systems allow for structuring of PCBs and advanced packaging for AI servers and optical interconnects.
19. $STM - their analog, power and microcontroller leadership powers edge-AI inference in humanoids, autos, and industrial IoT; strengthening pricing and diversified end-markets position it as Europe’s most resilient AI-chip winner.
20. $SMT (Scottish Mortgage Investment Trust) - ok, this one’s cheating a little lol. They’re a FTSE100 listed Trust (kinda like an ETF) with holdings in companies like $TSM, SpaceX, $NVDA, and $META.
A lot of these are AI/tech focussed = more fun. Lots of smaller MC names I've left off just due to the uncertainty.
Europe has a ton of huge names that won't see any growth over the next decade. In sectors like banking, manufacturing, automotives, pharmaceuticals.....very boring investments.
What just happened?
Yesterday, at 8:45 AM ET, Iran's Foreign Minister Araghchi announced that the Strait of Hormuz was "completely open" for all commercial vessels.
At 9:06 AM ET, President Trump thanked Iran for reopening the Strait.
Then, at 10:20 AM ET, Trump said Iran and the US were working together to remove all mines from the Strait of Hormuz.
Between 10:40 AM ET and 12:00 PM ET, President Trump said Iran agreed to "never close the Strait again" and to "suspend its nuclear program indefinitely."
Suddenly, at 6:14 PM ET, Iran's Speaker of the Parliament said Trump made "seven claims in one hour, all seven of which were false."
Now, Iran has CLOSED the Strait of Hormuz again and oil tankers are being struck.
What just happened behind the scenes?
Fed minutes show most Fed officials would cut rates into an negative oil shock due to its impact on employment. But a sizable minority is emphasizing the inflationary impacts and would instead hike rates.